[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10},"actionData","errors","root",{"_460":461,"_462":463},"routes/_app.x.$handle.$name",{"_110":109,"_38":55,"_118":100,"_33":119,"_13":14,"_120":-5,"_11":121,"_122":123,"_124":125,"_126":127,"_128":129,"_27":-5,"_130":-5,"_131":-5,"_132":31,"_133":134,"_135":69,"_136":-5,"_137":-5,"_138":139},"routes/_app.x.$handle.$name.branch.$branch.version.$version.diff",{"_11":12,"_13":14,"_15":16,"_17":18,"_19":-5},"project",{"_53":54,"_38":55,"_42":56,"_57":-5,"_58":31,"_59":60,"_61":60,"_62":63,"_64":-5,"_19":65,"_66":67,"_68":69,"_70":-5,"_71":-5,"_72":73,"_74":75,"_76":77,"_78":79,"_80":81,"_82":-5,"_83":84,"_85":84,"_86":87,"_88":87,"_89":87,"_90":84,"_91":84,"_92":93,"_94":95,"_96":97,"_98":31},"version",64,"changes",[25,26],"aiChatInfo",{"_20":21,"_22":-5,"_23":-5,"_24":-5},"userId","aiChatMessageId","754b7111-db2a-450f-98bb-73f2ddf36a61","aiChatThreadId","aiChatThreadUserId","aiChatThreadTitle",{"_27":47,"_29":-7,"_30":31,"_32":31,"_33":48,"_35":48},{"_27":28,"_29":-7,"_30":31,"_32":31,"_33":34,"_35":34},"file",{"_36":37,"_38":34,"_39":40,"_41":-5,"_42":43,"_44":14,"_45":46,"_30":31},"previous","deleted",false,"deletedHere","path","docs","latestPath","fileId","51e1e629-cf2e-47b6-b12a-3a334c07da3e","name","fileType","directory","content","createdAt",["D",1788495223471],"branchVersion","branchId","faea9245-6732-47d9-a1a2-3550998bc296",{"_36":49,"_38":50,"_39":27,"_41":51,"_42":52,"_44":14,"_45":46,"_30":31},"docs/growth-corridors.md","044917de-a817-11f1-9be5-1607ee4eb77e","growth-corridors.md","# DLX Designs · 50 Growth Corridors\n\nA professional framework for turning Nigerian marketplaces, fashion platforms,\ndirectories, and press into three things DLX actually needs:\n\n- **Acquisition** — post/sell where premium buyers browse.\n- **Observation** — watch where grooms, executives, and chieftaincies signal intent.\n- **Partnership** — list, feature, and collaborate with trusted local leaders.\n\nEvery corridor below maps to a **role the engine can play**, not a promise of\nany specific outcome. Use `source` (e.g. `jumia`, `bella-naija-weddings`,\n`nairaland`), route organic DMs/inquiries into the pipeline via `POST /api/ingest`\nor forward them to the email intake address, and let the qualifier + analytics\nshow which corridors actually produce qualified revenue.\n\n> **Honesty note:** verify each link/handle once before committing outreach —\n> marketplaces and directories change hands or go dormant. The ones marked\n> ⭐ major are large, long-running platforms. The rest are candidates to confirm.\n\n---\n\n## 1. General & Multi-Vendor Marketplaces — *Acquisition*\n| # | Corridor | DLX angle |\n|---|---|---|\n| 1 | **Jumia NG** ⭐ | Seller storefront + sponsored listings; Jumia's audience searches familiar Suits/Agbada |\n| 2 | **Konga** ⭐ | Multi-vendor storefront for premium categories |\n| 3 | **Jiji NG** ⭐ | Classifieds — high consideration, direct inquiry (feeds `/api/ingest`) |\n| 4 | **Temu NG** ⭐ | Benchmark price floors + observe category demand (not a premium fit) |\n| 5 | **PayPorte** | Storefront + style catalog for made-in-Nigeria fashion |\n| 6 | **Kusnap** | SME marketplace; Groomwear niche sellers |\n| 7 | **Sellevo** | Nigerian stores directory; local visibility |\n| 8 | **UPshop** | Local boutique discovery |\n| 9 | **AliExpress NG** ⭐ | Reference pricing/fabric categories; outbound compare |\n| 10 | **Nawao** | Commerce store; low volume but easy listing |\n\n## 2. Fashion, Apparel & Men's Style Platforms — *Acquisition + Observation*\n| # | Corridor | DLX angle |\n|---|---|---|\n| 11 | **Ajebomarket** | Men's fashion vertical listing |\n| 12 | **Garmspot** | Style store with buyer intent signals |\n| 13 | **Obeezi** | Men's wear focused — direct fit |\n| 14 | **The 5K Shop** | Retail portal for traditional wear |\n| 15 | **Miskay Boutique** | Boutique-style presentation for bespoke work |\n| 16 | **Sojoee** | Clothing store; test limited collections |\n| 17 | **Kaiglo** | E-commerce with Nigerian fashion categories |\n| 18 | **Bentigos** | Retail hub browse behavior observation |\n| 19 | **Mobos Fashion** | Apparel store breadth |\n| 20 | **Farfetch** ⭐ | Luxury positioning benchmark; not a volume channel |\n\n## 3. Directories & Artisan Networks — *Partnership + Acquisition*\n| # | Corridor | DLX angle |\n|---|---|---|\n| 21 | **BusinessList Lagos** | Premium listing under Fashion |\n| 22 | **TailorPro NG** | Artisan/made-to-measure directory — strong fit |\n| 23 | **VisCorner NG** | Tailoring (Lagos) — direct lead source |\n| 24 | **ArtisanOga** | Tailors directory; request/quote flow |\n| 25 | **Finelib Lagos** | Fashion designers index |\n| 26 | **Takooka** | Business portal (Fashion Lagos) |\n| 27 | **InfoisInfo Lagos** | Directory + review reputation |\n| 28 | **ConnectNigeria** ⭐ | Business listings, high SERP trust |\n| 29 | **VConnect** | Tailoring directory with callbacks |\n| 30 | **Sellatease** | Marketplace directory for exporters/sellers |\n\n## 4. Fashion Week & Industry Hubs — *Partnership + Authority*\n| # | Corridor | DLX angle |\n|---|---|---|\n| 31 | **Lagos Fashion Week** ⭐ | Showcase/network; authority association |\n| 32 | **Africa Fashion Week NG** | Designer participant/partner |\n| 33 | **Printivo** | Creative/merch partner for collateral |\n| 34 | **StyleHouse Files** | Editorial creative directory |\n| 35 | **NTMA** | Fabric sourcing ecosystem (partner for premium materials) |\n\n## 5. Classifieds, Forums & Community Boards — *Observation + Acquisition*\n| # | Corridor | DLX angle |\n|---|---|---|\n| 36 | **Nairaland — Business** ⭐ | Large Nigerian forum; watch demand, reply to genuine seekers |\n| 37 | **Nairaland — Fashion** ⭐ | Groom/agbada threads = real intent signals |\n| 38 | **Ogun State Business Directory** | Abeokuta regional coverage |\n| 39 | **Lagos MSME/Enterprise Portal** | Government-backed visibility + grants ecosystem |\n| 40 | **MAN** | Corporate/executive B2B channel |\n\n## 6. Lifestyle, Luxury & Executive Press — *Authority + Observation*\n| # | Corridor | DLX angle |\n|---|---|---|\n| 41 | **BellaNaija Weddings** ⭐ | Groom features → wedding-intent audience |\n| 42 | **ThisDay Style** ⭐ | Luxury menswear editorial placement |\n| 43 | **BusinessDay Lifestyle** ⭐ | Executive readership outreach |\n| 44 | **Guardian Life & Style** ⭐ | Prestige + credibility |\n| 45 | **Vanguard Allure** ⭐ | Style vertical reach |\n| 46 | **Nairametrics** ⭐ | Business/exec audience alignment |\n| 47 | **Pulse NG — Fashion** ⭐ | Cultural relevance + trends |\n| 48 | **Style Vitae** | Local style editorial |\n| 49 | **Eventbrite Lagos** ⭐ | Fashion/business events to attend or sponsor |\n| 50 | **Meetup Lagos / Groups** ⭐ | Executive & fashion networking clusters |\n\n---\n\n## How the engine consumes each corridor\n\n1. **Route the inquiry** — every DM, form, or email from a corridor lands tagged\n with the corridor as `source`.\n2. **Qualify** — the strict engine accepts only genuine high-intent (≥ 50),\n rejects budget seekers, and skips duplicates.\n3. **Prioritize** — the dashboard surfaces strongest opportunities first.\n4. **Track & learn** — analytics report **which corridors and signals actually\n convert through the pipeline to deposit/ordered/won**, so budget flows to what\n produces revenue — not to vanity listings.\n\nStart with the ⭐ majors (Jumia, Konga, Jiji, ConnectNigeria, Nairaland,\nBellaNaija Weddings, Lagos Fashion Week, the national press) — they carry the\nmost real, high-intent audience for premium bespoke menswear today.\n",["D",1788495223471],"id","d2f8de76-3d02-4949-8a06-2f7cc759dae3","dlx-lead-engine",["D",1788452396724],"deletedAt","legacyEnvVars","privacy","public","httpPrivacy","description","DLX Designs lead qualification engine + React dashboard","imageUrl","0e434874-725d-4e49-bc72-a0c364fc943f","pinnedFileId","01a06811-df40-730d-85ee-1ae96a373bd7","pinPreview",true,"pinMigrated","screenshotVersion","likeCount",0,"orgId","01a04e8e-cc2d-73ad-9a1b-036254519f63","tursoName","d2f8de763d0249498a062f7cc759dae3","tursoHostname","d2f8de763d0249498a062f7cc759dae3-val-town.aws-us-east-2.turso.io","tursoMeasuredAt",["D",1789494036282],"pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_110":109,"_117":-5},"org",{"_110":109,"_111":112,"_113":-5,"_19":-5,"_114":115,"_116":31},"branch",{"_13":99,"_38":100,"_53":46,"_101":54,"_102":-5,"_103":-5,"_104":-5,"_105":-5,"_106":-5,"_107":55,"_108":109},"isOwned",88,"main","projectId","forkedBranchId","forkedBranchVersion","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","dlxdesigns","handle","displayName","DLXdesigns ","avatar","tier","free","canRestrictVals","profileImageUrl","branchName","","adoptOrgId",{"_53":54,"_38":55,"_42":455,"_57":-5,"_58":31,"_59":60,"_61":60,"_62":63,"_64":-5,"_19":65,"_66":67,"_68":69,"_70":-5,"_71":-5,"_72":73,"_74":75,"_76":77,"_78":79,"_80":456,"_82":-5,"_83":84,"_85":84,"_86":87,"_88":87,"_89":87,"_90":84,"_91":84,"_92":457,"_94":458,"_96":459,"_98":31},"tree",[232,239,241,243,245,247,256,258,261,262,263],"files",[232,233,234,235,236,237,238,239,240,241,242,243,244,245,246,247,248,249,250,251,252,253,254,255,256,257,258,259,260,261,262,263],"counts",{"_222":73,"_223":174,"_224":73,"_225":226,"_227":228},"contents",[140,141,142,143,144,145,146,147,148,149,150],"upstreamProject","upstreamBranch","emptyEnvVars","flags",[],"hasHttpFiles","grants","orgs","customDomains",{},{"_53":216,"_38":217,"_36":218,"_153":-5,"_45":46,"_44":219,"_155":69,"_42":220,"_13":73,"_39":40,"_41":-5,"_30":31,"_160":-5,"_33":221},{"_53":210,"_38":211,"_36":212,"_153":-5,"_45":46,"_44":213,"_155":69,"_42":214,"_13":73,"_39":40,"_41":-5,"_30":31,"_160":-5,"_33":215},{"_53":204,"_38":205,"_36":206,"_153":-5,"_45":46,"_44":207,"_155":69,"_42":208,"_13":73,"_39":40,"_41":-5,"_30":31,"_160":-5,"_33":209},{"_53":201,"_38":34,"_36":37,"_153":-5,"_45":46,"_44":14,"_155":69,"_42":202,"_13":73,"_39":40,"_41":-5,"_30":31,"_160":-5,"_33":203},{"_53":195,"_38":196,"_36":197,"_153":-5,"_45":46,"_44":198,"_155":69,"_42":199,"_13":73,"_39":40,"_41":-5,"_30":31,"_160":-5,"_33":200},{"_53":190,"_38":191,"_36":192,"_153":-5,"_45":46,"_44":174,"_155":69,"_42":193,"_13":73,"_39":40,"_41":-5,"_30":31,"_160":-5,"_33":194},{"_53":184,"_38":185,"_36":186,"_153":-5,"_45":46,"_44":187,"_155":69,"_42":188,"_13":73,"_39":40,"_41":-5,"_30":31,"_160":-5,"_33":189},{"_53":178,"_38":179,"_36":180,"_153":-5,"_45":46,"_44":181,"_155":69,"_42":182,"_13":73,"_39":40,"_41":-5,"_30":31,"_160":-5,"_33":183},{"_53":171,"_38":172,"_36":173,"_153":-5,"_45":46,"_44":174,"_155":69,"_42":175,"_13":73,"_39":27,"_41":176,"_30":31,"_160":-5,"_33":177},{"_53":163,"_38":164,"_36":165,"_153":-5,"_45":46,"_44":166,"_155":31,"_42":167,"_13":168,"_39":27,"_41":169,"_30":31,"_160":-5,"_33":170},{"_53":151,"_38":152,"_36":67,"_153":-5,"_45":46,"_44":154,"_155":31,"_42":156,"_13":157,"_39":158,"_41":159,"_30":31,"_160":161,"_33":162},"dccbdd47-020c-4a20-86ae-ef8c8a971747","index.ts","parentId",57,"latest",["D",1788494730507],5,"http","import { parseVal, serveFile, serveImmutableFile } from \"https://esm.town/v/std/utils/index.ts\";\nimport { Hono } from \"npm:hono\";\nimport { Root } from \"./frontend/root.tsx\";\nimport {\n listLeads,\n setLeadStatus,\n stageDistribution,\n topSources,\n topSignals,\n} from \"./lib/store.ts\";\nimport { ingestLead } from \"./backend/worker/ingest.ts\";\nimport { nextStage } from \"./backend/pipeline.ts\";\n\nconst app = new Hono();\n\n// Serve the (never-cached) HTML shell at the root and as the dashboard\napp.get(\"/\", (c) => c.html(Root()));\napp.get(\"/dashboard\", (c) => c.html(Root()));\n\n// Standalone, embeddable lead-capture form (Option A: website ingestion)\napp.get(\"/capture\", (c) => serveFile(\"/frontend/capture.html\"));\n\n// Serve frontend files at the version-addressed URLs Root() stamps, cached immutably\napp.get(\"/__immutable/*\", (c) => serveImmutableFile(c.req.path));\n\n// View source redirect\napp.get(\"/source\", (c) => c.redirect(parseVal().links.self.val));\n\n// --- Lead API -----------------------------------------------------------\n\n// GET verified (already-qualified) leads for the dashboard\napp.get(\"/api/leads\", async (c) => {\n const leads = await listLeads(100);\n return c.json(leads);\n});\n\n// POST an inbound lead: { name, gmail, message, source }\n// Runs the strict qualification engine. Only high-intent leads are stored;\n// budget seekers and duplicates are rejected (422) and never persisted.\napp.post(\"/api/ingest\", async (c) => {\n const body = await c.req.json().catch(() => undefined);\n const outcome = await ingestLead((body as object) ?? {});\n return c.json(outcome, outcome.decision === \"reject\" ? 422 : 201);\n});\n\n// Move a lead forward in the pipeline (or set any valid stage).\n// Automatically advances when no stage body is given (\"next\").\napp.post(\"/api/leads/:id/status\", async (c) => {\n const id = c.req.param(\"id\");\n const body = (await c.req.json().catch(() => undefined)) as\n | { stage?: unknown; note?: unknown }\n | undefined;\n const lead = await listLeads(200);\n const existing = lead.find((l) => l.id === id);\n if (!existing) return c.json({ error: \"lead not found\" }, 404);\n\n let target: string;\n if (body?.stage) {\n target = String(body.stage);\n } else {\n const nxt = nextStage(existing.status as never);\n if (!nxt) return c.json({ error: \"terminal stage\", lead: existing });\n target = nxt;\n }\n\n const updated = await setLeadStatus(id, target, String(body?.note ?? \"\"));\n return c.json(updated ?? { error: \"lead not found\" }, updated ? 200 : 404);\n});\n\n// Analytics — what actually converts (learn layer).\napp.get(\"/api/analytics\", async (c) => {\n const [stages, sources, signals] = await Promise.all([\n stageDistribution(),\n topSources(8),\n topSignals(8),\n ]);\n return c.json({ stages, sources, signals });\n});\n\n// Unwrap and rethrow Hono errors as the original error\n// If you delete this, all errors will be \"Internal Server Error\"\napp.onError((err) => Promise.reject(err));\n\n// HTTP vals expect an exported \"fetch handler\"\n// This is how you \"run the server\" in Val Town with Hono\nexport default app.fetch;\n","label","dlx-leads",[152],"6d8419b9-a5f9-46c1-a394-d0a65ac33ae6","README.md","01a06811-df40-730d-85ee-161b7548ebf9",60,["D",1788494749875],17,"# DLX Designs · Verified Lead Intelligence Engine\n\nProduction-ready lead-generation and **strict qualification** engine for DLX\nDesigns — a luxury bespoke traditional menswear brand targeting high-net-worth\nclients, grooms and corporate executives in Lagos & Abeokuta, Nigeria.\n\nIt ingests real inbound traffic, filters out low-budget seekers with strict\nsemantic rules, stores only verified high-intent leads, and renders them as\nscored lead cards in a live dashboard.\n\n> **Live dashboard:** [Try the app](url:index.ts)\n> **Ingest endpoint:** `POST /api/ingest`\n\n## Architecture\n\n```mermaid\ngraph LR\n Inbound[\"Web form / corridor
(blog, IG, ads)\"] -->|\"POST /api/ingest\"| Qual[\"Qualifier
backend/worker/qualifier.ts\"]\n Qual -->|\"hard reject\"| X[\"Low budget → dropped
never stored\"]\n Qual -->|\"optional\"| LLM[\"AI refine
(free tier, optional)\"]\n Qual -->|\"pass ≥ 50\"| Store[\"SQLite store
lib/store.ts\"]\n Store -->|\"optional mirror\"| Supabase[\"Supabase
dlx_verified_leads\"]\n Store -->|\"GET /api/leads\"| UI[\"React dashboard
frontend/\"]\n```\n\n- **Dashboard:** client-side React + Hono backend (App Router-equivalent), Twind/Tailwind styling.\n- **Database:** val-scoped PostgreSQL-compatible SQLite. **Supabase mirror optional.**\n- **Qualifier:** deterministic heuristic (zero-config) + optional free-tier AI refinement.\n\n## The strict qualification engine\n\nRuns in `backend/worker/qualifier.ts` and never fabricates data.\n\n**Layer 1 — hard disqualifiers (instant reject):** `cheap`, `affordable`,\n`under 20k`/`20k`, `discount`, `thrift`, `okirika`.\n\n**Layer 2 — scoring matrix (0–100):** baseline 25, high-tier signals add points —\n`bespoke` +18, `groom`/`chieftaincy` +20, `cashmere` +15, `wedding` +14,\n`3-piece` +12, `executive` +14, `lekki`/`victoria island` +10, `lagos`/`abeokuta`\n+6, etc. **Pass threshold: ≥ 50.** Sub-threshold leads are rejected.\n\n**Layer 3 — AI refinement (optional, zero-config by default):** the engine is\n**deterministic-first** — the strict heuristic above is authoritative and needs\n**no AI key, no external call, no cost**, so no provider can ever block it. If\nyou add a free-tier key, that provider refines the score and writes a natural\nbuying signal (tried in order: `GEMINI_API_KEY`, `GROQ_API_KEY`, `OPENAI_API_KEY`).\nA provider failing just falls back to the heuristic — never an error.\n\n### Endpoints\n\n| Method | Path | Purpose |\n| --- | --- | --- |\n| `GET` | `/` | Dashboard |\n| `GET` | `/dashboard` | Dashboard (alias) |\n| `GET` | `/api/leads` | Verified leads (JSON, strongest first) |\n| `POST` | `/api/ingest` | Ingest & qualify an inbound lead — 201 on pass, 422 on reject |\n| `POST` | `/api/leads/:id/status` | Advance pipeline stage (or set one) — records history |\n| `GET` | `/api/analytics` | Pipeline / source / signal analytics (the learn layer) |\n\n`POST /api/ingest` request body:\n\n```json\n{\n \"name\": \"Oluwatosin Adeyemi\",\n \"gmail\": \"tosin@example.com\",\n \"source\": \"nigerianmenswear.com\",\n \"message\": \"I need a bespoke 3-piece groom suit for my chieftaincy ceremony in Abeokuta\"\n}\n```\n\nOnly passing leads persist. Rejected and duplicate leads are returned (422) and\nnever stored.\n\n## The growth engine loop\n\nA verified lead is more than a row — it is a story with evidence:\n\n1. **Hunt** — `cron/hunt.ts` polls a real configured source (`DLX_HUNT_URL`)\n and feeds genuinely captured inquiries through the qualifier. It never\n invents people or signals; rows without evidence are ignored.\n2. **Verify** — every lead records where it came from (`source`, `source_url`),\n exactly what was said (`evidence`), and when (`created_at`).\n3. **Qualify** — strict scoring (0–100); only ≥ 50 passes.\n4. **Prioritize** — the dashboard sorts strongest first and shows live pipeline counts.\n5. **Recommend** — each lead carries a `action`, chosen from the actual signal:\n book consultation, contact to confirm budget, or nurture.\n6. **Track** — leads move through `new → qualified → contacted → responded →\n consultation → quoted → deposit_paid → ordered → won/lost`. Every transition\n is recorded in `history`.\n7. **Learn** — `/api/analytics` reports conversion by stage and top sources and\n signals that actually produce qualified opportunities.\n\n## Capture & feed real inquiries (get it live today)\n\n**Option A — website ingestion.** A ready, branded enquiry form is hosted at\n`/capture` (or embedded by iframing it). It posts each submission to\n`/api/ingest`, which qualifies and stores it. Direct API for a custom site:\n\n```http\nPOST https://dlx-leads.val.run/api/ingest\nContent-Type: application/json\n\n{ \"name\": \"…\", \"gmail\": \"…\", \"phone\": \"…\", \"message\": \"…\", \"source\": \"…\", \"source_url\": \"…\" }\n```\n\n**Option C — JSON / Google Sheet feed (autonomous, scheduled every 15 min).**\nThe hunt poller accepts a **published Google Sheet CSV** directly, no code:\n\n1. In your sheet, make row 1 a header that includes `gmail` (plus optional\n `name`, `phone`, `message`, `source`, `source_url`).\n2. **File → Share → Publish to web →** choose the sheet → format **CSV** → copy the link.\n3. Set that link as `DLX_HUNT_URL` (and `DLX_HUNT_TOKEN` if it's private/behind auth).\n\nEach run, the engine ingests every new row, qualifies it, and skips duplicates.\nRows without an email are ignored — nothing is invented.\n\n## Configuration (environment variables)\n\nEverything runs out of the box with **zero environment variables**. The\nqualification engine is fully deterministic. The following optional keys unlock\nthe AI refinement layer (free tiers are preferred — none require billing) and\nthe Supabase mirror:\n\n- **`GEMINI_API_KEY`** — free tier (Google AI Studio key, no card required). Recommended.\n- **`GROQ_API_KEY`** — free tier. Ignored unless Gemini is absent or fails.\n- **`OPENAI_API_KEY`** — requires active OpenAI billing; only used as a last resort.\n- **`SUPABASE_URL`** + **`SUPABASE_SERVICE_ROLE_KEY`** — mirror verified leads to\n your Supabase `dlx_verified_leads` table (native SQLite store stays authoritative).\n- **`DLX_HUNT_URL`** — JSON **or published-CSV** feed the hunt poller watches\n for genuine inquiries (JSON: `{ \"leads\": [ { name, gmail, phone, message, source, source_url } ] }`).\n- **`DLX_HUNT_TOKEN`** — optional bearer token sent to the hunt source (`Authorization: Bearer …`).\n\n👉 Add GEMINI_API_KEY here: https://www.val.town/x/dlxdesigns/dlx-lead-engine/environment-variables?key=GEMINI_API_KEY\n👉 Add GROQ_API_KEY here: https://www.val.town/x/dlxdesigns/dlx-lead-engine/environment-variables?key=GROQ_API_KEY\n👉 Add OPENAI_API_KEY here: https://www.val.town/x/dlxdesigns/dlx-lead-engine/environment-variables?key=OPENAI_API_KEY\n👉 Add SUPABASE_URL here: https://www.val.town/x/dlxdesigns/dlx-lead-engine/environment-variables?key=SUPABASE_URL\n👉 Add SUPABASE_SERVICE_ROLE_KEY here: https://www.val.town/x/dlxdesigns/dlx-lead-engine/environment-variables?key=SUPABASE_SERVICE_ROLE_KEY\n👉 Add DLX_HUNT_URL here: https://www.val.town/x/dlxdesigns/dlx-lead-engine/environment-variables?key=DLX_HUNT_URL\n👉 Add DLX_HUNT_TOKEN here: https://www.val.town/x/dlxdesigns/dlx-lead-engine/environment-variables?key=DLX_HUNT_TOKEN\n\n### Supabase setup (optional)\n\nCreate the mirror table by running [`database/schema.sql`](database/schema.sql)\nin the Supabase SQL editor. The table uses `gen_random_uuid()` ids, enforces a\n`score BETWEEN 0 AND 100` check, has a `gmail UNIQUE` constraint, and enables\nRow-Level Security (locked to service-role access by default). The val writes\nvia PostgREST with the service-role key and `resolution=merge-duplicates`.\n\n## File structure\n\n```\ndatabase/\n schema.sql ← Supabase/PostgreSQL schema (mirror table)\nbackend/\n pipeline.ts ← sales pipeline stages + next-stage logic\n hunt.ts ← autonomous ingestion from a real configured source\n worker/\n qualifier.ts ← strict qualification engine (disqualifiers + scoring)\n ingest.ts ← ingest worker (bridge inbound → store)\n llm.ts ← optional AI adapter (free tier Gemini / Groq / OpenAI)\ncron/\n hunt.ts ← scheduled hunt poller (every 15 min)\nlib/\n store.ts ← SQLite persistence + Supabase mirror + pipeline + analytics\nfrontend/\n components/App.tsx ← dashboard layout + empty state + pipeline strip\n components/LeadCard.tsx ← verified-lead card with evidence, action & stage\n components/LeadForm.tsx ← inline lead-capture form\nscripts/\n selftest.ts ← end-to-end qualifier + store self-test\nindex.ts ← Hono HTTP backend (dashboard + /api/*)\n```\n\n## Integrity guarantees\n\n- **No hardcoded, fake, or mock contacts anywhere.** Every record comes from a\n real `POST /api/ingest` submission that passed strict qualification.\n- Empty store → dashboard shows *“Awaiting live inbound high-intent signals…”*,\n never fake cards.\n- Duplicate emails are rejected (UNIQUE), budget seekers are dropped, sub-50\n intent is dropped.\n\n## Running the self-test\n\n`scripts/selftest.ts` exercises the full pipeline with synthetic example inputs\n(high-intent pass, low-budget reject, no-signal reject, duplicate reject) and is\nsafe to run repeatedly.\n",[164],"a130fa68-e89b-4e79-b351-ae119f51052e","CHANGELOG.md","01a06811-df40-730d-85ee-12df63d9ba38",1,["D",1788452396724],"# Changelog\n\n## 2026-07-30 — Immutable asset caching\n\nFrontend files are now served at version-addressed URLs\n(`/__immutable/45/frontend/index.tsx`) with\n`Cache-Control: public, max-age=31536000, immutable`, so repeat visits load the\nwhole client graph from the browser cache — roughly 3× faster (~665ms → ~157ms\nmeasured). Publishing your val bumps its version, which changes every URL, so\ncaches invalidate automatically. Old-version URLs return 404 after a publish\n(like Next.js build assets); a reload picks up the new version.\n\nWhat changed in the template:\n\n- `frontend/index.html` → [`frontend/root.tsx`](/frontend/root.tsx): the HTML\n shell is now JSX and stamps asset URLs with `immutableFileUrl()`\n- [`index.ts`](/index.ts): serves `Root()` at `/`, and one\n `app.get(\"/__immutable/*\", (c) => serveImmutableFile(c.req.path))` route\n serves all frontend files\n\n### Upgrading an existing remix\n\nYour copy of this template doesn't update automatically. Two options:\n\n**Minimal (most of the benefit):** point both routes at `serveImmutableFile` —\nyour existing bare URLs redirect into versioned space, the new route serves\nthem, and the whole module graph gets cached (costs one redirect per page view):\n\n```ts\nimport { serveImmutableFile } from \"https://esm.town/v/std/utils/index.ts\";\n\napp.get(\"/__immutable/*\", (c) => serveImmutableFile(c.req.path));\napp.get(\"/frontend/**/*\", (c) => serveImmutableFile(c.req.path)); // was serveFile\n```\n\n**Full upgrade (also removes the redirect):** three self-contained steps —\nno other files needed. Adjust the ``, favicon, and body of `Root()` to\nmatch your app's existing `index.html`, and stamp anything else your shell\nreferences directly (images, css) with `immutableFileUrl()`.\n\n1. Create `frontend/root.tsx`:\n\n```tsx\n/** @jsxImportSource npm:hono@4/jsx */\nimport { raw } from \"npm:hono@4/html\";\nimport { immutableFileUrl } from \"https://esm.town/v/std/utils/index.ts\";\n\n// The HTML shell for the app. It is never cached, so immutableFileUrl can stamp\n// the current val version into each asset URL — the browser caches those\n// immutably, and publishing the val changes the URLs, refetching exactly once.\nexport function Root() {\n return (\n <>\n {raw(\"<!DOCTYPE html>\")}\n <html lang=\"en\">\n <head>\n <meta charSet=\"UTF-8\" />\n <meta name=\"viewport\" content=\"width=device-width, initial-scale=1.0\" />\n <title>React Hono Val Town Starter\n \n \n \n \n

DLX Designs

Bespoke menswear enquiry

Tell us what you need. Verified inquiries are handled personally.

\n\n \n\n

DLX Designs · Lagos & Abeokuta

\n\n \n \n\n",[],[191,327],"909f573f-b663-4da1-9e37-58f6e7fbf8fb","LeadForm.tsx","bd44cf7c-a7b7-11f1-bb36-1607ee4eb77e","01a06811-df3f-7558-9972-4e3e2fdb144e",16,["D",1788454302131],"/** @jsxImportSource https://esm.sh/react@18.2.0 */\n// frontend/components/LeadForm.tsx\n// Inline lead-capture form. Posts to /api/ingest, which runs the strict\n// qualifier. Shows pass/reject feedback inline.\nimport { useState, type FormEvent } from \"https://esm.sh/react@18.2.0\";\n\ntype SubmitState =\n | { kind: \"idle\" }\n | { kind: \"sending\" }\n | { kind: \"ok\"; message: string }\n | { kind: \"err\"; message: string };\n\nexport function LeadForm({ onCaptured }: { onCaptured: () => void }) {\n const [name, setName] = useState(\"\");\n const [gmail, setGmail] = useState(\"\");\n const [source, setSource] = useState(\"\");\n const [message, setMessage] = useState(\"\");\n const [state, setState] = useState({ kind: \"idle\" });\n\n async function submit(e: FormEvent) {\n e.preventDefault();\n setState({ kind: \"sending\" });\n try {\n const res = await fetch(\"/api/ingest\", {\n method: \"POST\",\n headers: { \"Content-Type\": \"application/json\" },\n body: JSON.stringify({ name, gmail, source, message }),\n });\n const data = await res.json();\n if (data.decision === \"pass\") {\n setState({ kind: \"ok\", message: `Verified lead stored · ${data.lead?.score}` });\n setName(\"\");\n setGmail(\"\");\n setSource(\"\");\n setMessage(\"\");\n onCaptured();\n } else {\n setState({ kind: \"err\", message: `Not qualified: ${data.reason ?? \"rejected\"}` });\n }\n } catch (err) {\n setState({ kind: \"err\", message: String(err instanceof Error ? err.message : err) });\n }\n }\n\n const inputCls =\n \"w-full rounded-lg border border-stone-300 bg-white px-3 py-2 text-sm text-stone-800 placeholder:text-stone-400 focus:outline-none focus:ring-2 focus:ring-amber-500\";\n\n return (\n

Capture an inbound lead

\n POST /api/ingest\n

void submit(e)} class=\"grid grid-cols-1 sm:grid-cols-2 gap-3\">\n setName(e.target.value)}\n placeholder=\"Client name\"\n required\n />\n setGmail(e.target.value)}\n placeholder=\"Email address\"\n required\n />\n setSource(e.target.value)}\n placeholder=\"Source / domain (e.g. nigerianmenswear.com)\"\n />\n setMessage(e.target.value)}\n placeholder='What they asked for, e.g. \"bespoke 3-piece groom suit\"'\n required\n />\n

\n {state.kind === \"ok\" && (\n {state.message}\n )}\n {state.kind === \"err\" && (\n {state.message}\n )}\n \n {state.kind === \"sending\" ? \"Qualifying…\" : \"Capture lead\"}\n \n

\n \n

\n );\n}\n",[],[191,343,335],"components","f016accd-d506-4618-b8dd-f0f37419a6b2","LeadCard.tsx","825db1fc-a7b3-11f1-a292-1607ee4eb77e",40,["D",1788493743320],"/** @jsxImportSource https://esm.sh/react@18.2.0 */\n// frontend/components/LeadCard.tsx\n// A single verified-lead card: identity, evidence, signal, recommended action,\n// and pipeline stage controls. Nothing here is ever invented — it renders only\n// what the store captured.\nimport { useState } from \"https://esm.sh/react@18.2.0\";\n\nexport interface StageEvent {\n from: string;\n to: string;\n at: string;\n note?: string;\n}\n\nexport interface Lead {\n id: string;\n name: string;\n gmail: string;\n phone: string;\n score: number;\n source: string;\n source_url: string;\n evidence: string;\n signal: string;\n action: string;\n status: string;\n history: StageEvent[];\n created_at: string;\n updated_at: string;\n}\n\nconst STAGE_LABELS: Record = {\n new: \"New\", qualified: \"Qualified\", contacted: \"Contacted\",\n responded: \"Responded\", consultation: \"Consultation\", quoted: \"Quoted\",\n deposit_paid: \"Deposit Paid\", ordered: \"Ordered\", won: \"Won\", lost: \"Lost\",\n};\n\nfunction scoreNumber(n: number): number {\n return Number.isFinite(n) ? Math.min(100, Math.max(0, Math.round(n))) : 0;\n}\n\nfunction badgeClass(score: number): string {\n const s = scoreNumber(score);\n if (s >= 80) return \"bg-emerald-100 text-emerald-800 border-emerald-300\";\n if (s >= 65) return \"bg-teal-100 text-teal-800 border-teal-300\";\n if (s >= 50) return \"bg-amber-100 text-amber-800 border-amber-300\";\n return \"bg-gray-100 text-gray-600 border-gray-300\";\n}\n\nfunction initials(name: string): string {\n return name.split(/\\s+/).filter(Boolean).slice(0, 2)\n .map((p) => p[0]?.toUpperCase() ?? \"\").join(\"\") || \"DLX\";\n}\n\nexport function LeadCard({ lead, onAdvance }: { lead: Lead; onAdvance: (id: string) => void }) {\n const s = scoreNumber(lead.score);\n const terminal = lead.status === \"won\" || lead.status === \"lost\";\n const [busy, setBusy] = useState(false);\n const date = new Date(lead.created_at);\n const when = Number.isNaN(date.getTime())\n ? \"\"\n : date.toLocaleDateString(undefined, { day: \"numeric\", month: \"short\" });\n\n return (\n

\n {/* Identity + score */}\n

\n {initials(lead.name)}\n

{lead.name}

{lead.gmail}

\n {lead.phone && (\n

📱 {lead.phone}

\n )}\n

\n \n {s}/100%\n \n

\n\n {/* Source + stage */}\n

\n \n {lead.source_url ? (\n \n 🌐 {lead.source || \"link\"}\n \n ) : (\n <>🌐 {lead.source || \"inbound-form\"}\n )}\n \n {when && (\n {when}\n )}\n \n {STAGE_LABELS[lead.status] ?? lead.status}\n \n

\n\n {/* Evidence — what was actually said */}\n {lead.evidence && (\n

\n Evidence\n

{lead.evidence}

\n )}\n\n {/* Signal + recommended action */}\n

Signal: {lead.signal || \"High-intent luxury menswear inquiry\"}

\n {lead.action && (\n

\n Next: {lead.action}\n

\n )}\n

\n\n {/* Pipeline control */}\n {!terminal ? (\n {\n setBusy(true);\n onAdvance(lead.id);\n setTimeout(() => setBusy(false), 800);\n }}\n disabled={busy}\n class=\"mt-1 rounded-lg border border-stone-300 hover:border-stone-900 hover:bg-stone-900 hover:text-white px-3 py-1.5 text-sm font-medium transition-colors cursor-pointer disabled:opacity-50\"\n >\n {busy ? \"Advancing…\" : \"Advance stage →\"}\n \n ) : (\n \n {lead.status === \"won\" ? \"🏆 Won\" : \"Closed / lost\"}\n \n )}\n

\n );\n}\n",[],[191,343,345],"6cdc14cb-cb85-4b41-8d69-c873f731c699","App.tsx","01a06811-df3f-7558-9972-46f66ec4eab9",45,["D",1788493768801],8,"/** @jsxImportSource https://esm.sh/react@18.2.0 */\n// frontend/components/App.tsx\n// DLX Designs — verified lead intelligence dashboard.\n// Zero placeholder data: renders only real records from the val's store via\n// /api/leads. Empty store shows a clean \"awaiting live inbound signals\" state.\nimport { useEffect, useState } from \"https://esm.sh/react@18.2.0\";\nimport { LeadCard, type Lead } from \"./LeadCard.tsx\";\nimport { LeadForm } from \"./LeadForm.tsx\";\n\nconst ORDER = [\n [\"new\", \"New\"], [\"qualified\", \"Qualified\"], [\"contacted\", \"Contacted\"],\n [\"responded\", \"Responded\"], [\"consultation\", \"Consultation\"], [\"quoted\", \"Quoted\"],\n [\"deposit_paid\", \"Deposit\"], [\"ordered\", \"Ordered\"], [\"won\", \"Won\"], [\"lost\", \"Lost\"],\n] as const;\n\nfunction EmptyState() {\n return (\n

\n 🧵\n

\n Awaiting live inbound high-intent signals…\n

\n Verified leads captured from your web forms and inbound corridors will\n appear here as scored cards. No placeholder records are shown.\n

\n POST /api/ingest {\"{ name, gmail, message, source }\"}\n

\n );\n}\n\nexport function App() {\n const [leads, setLeads] = useState(null);\n const [error, setError] = useState(null);\n const [loading, setLoading] = useState(true);\n\n async function refresh() {\n setLoading(true);\n try {\n const res = await fetch(\"/api/leads\");\n if (!res.ok) throw new Error(`API ${res.status}`);\n const data = (await res.json()) as Lead[];\n setLeads(Array.isArray(data) ? data : []);\n setError(null);\n } catch (e) {\n setError(String(e instanceof Error ? e.message : e));\n } finally {\n setLoading(false);\n }\n }\n\n useEffect(() => {\n void refresh();\n void loadAnalytics();\n }, []);\n\n const count = leads?.length ?? 0;\n const avg =\n leads && leads.length > 0\n ? Math.round(leads.reduce((a, l) => a + (l.score || 0), 0) / leads.length)\n : 0;\n const top = leads?.slice()\n .sort((a, b) => (b.score || 0) - (a.score || 0))[0];\n\n const [stages, setStages] = useState | null>(null);\n\n async function advance(id: string) {\n try {\n await fetch(`/api/leads/${id}/status`, {\n method: \"POST\",\n headers: { \"Content-Type\": \"application/json\" },\n body: \"{}\",\n });\n } catch {\n /* keep going */\n }\n await refresh();\n void loadAnalytics();\n }\n\n async function loadAnalytics() {\n try {\n const r = await fetch(\"/api/analytics\");\n const d = await r.json();\n const m: Record = {};\n for (const s of d?.stages ?? []) m[s.status] = s.count;\n setStages(m);\n } catch {\n /* optional layer */\n }\n }\n\n return (\n

\n DLX Designs\n

\n Verified Lead Intelligence\n

\n High-intent luxury menswear inquiries · Lagos & Abeokuta\n

\n void refresh()}\n class=\"inline-flex items-center gap-2 rounded-lg bg-amber-500 hover:bg-amber-400 text-stone-900 font-semibold px-4 py-2 text-sm transition-colors cursor-pointer disabled:opacity-50\"\n disabled={loading}\n >\n {loading ? \"Refreshing…\" : \"⟳ Refresh\"}\n \n

\n\n

\n {stages && (\n

\n Pipeline\n

\n {ORDER.map(([key, label]) => (\n 0\n ? \"bg-stone-900 text-stone-50 border-stone-900\"\n : \"bg-white text-stone-400 border-stone-200\"\n }`}\n >\n {label}: {stages[key] ?? 0}\n \n ))}\n

\n )}\n\n void refresh()} />\n\n {leads && leads.length > 0 && (\n

Verified leads

{count}

Avg intent score

{avg}/100

Hottest lead

\n {top ? `${top.name} · ${top.score ?? 0}/100` : \"—\"}\n

\n )}\n\n {error && (\n

\n Could not load leads: {error}\n

\n )}\n\n {leads === null && !error ? (\n

Loading verified leads…

\n ) : !error && count === 0 ? (\n \n ) : (\n

\n {leads?.map((l) => (\n void advance(id)} />\n ))}\n

\n )}\n\n \n

\n );\n}\n",[],[191,343,353],"c1571bc2-1eae-4a07-9531-058b3b6dd68d",["D",1788452396724],[249,250,251],[191,343],["D",1788452396724],[248,252,253,254,255],[191],"01029356-9c7a-410f-af74-6417c61c9a0c","lead-intake.ts","e7b272d2-a816-11f1-9be5-1607ee4eb77e",["D",1788495175441],"// email/lead-intake.ts\n// EMAIL INTAKE — the live inbound-hunting channel.\n//\n// Any email sent (or forwarded) to this val's address is run through the strict\n// qualifier and, if it passes, becomes a verified lead in the pipeline. This is\n// the closest thing to real-time inbound hunting: forward a DM/WhatsApp/web\n// inquiry to the address and it is qualified, evidence-tagged, and tracked\n// automatically — no spreadsheets, no forms.\n//\n// Find the assigned address in `links.email` on this file.\nimport { ingestLead } from \"../backend/worker/ingest.ts\";\n\nexport default async function (e: Email) {\n const from = e.from || \"\";\n const text = `${e.subject ?? \"\"}\\n\\n${e.text ?? \"\"}`.trim();\n\n // From: \"Ada Obi\" -> name + email\n const addressMatch = from.match(/<([^<>@\\s]+@[^<>@\\s]+)>/i);\n const bareEmail = (from.match(/\\S+@\\S+/) ?? [\"\"])[0];\n const gmail = (addressMatch?.[1] ?? bareEmail ?? \"\").toLowerCase().trim();\n const nameRaw = (from.split(\"<\")[0] || \"\")\n .replace(/[\"']/g, \"\")\n .replace(/^\\s*(From|mailto)?:?\\s*/i, \"\")\n .trim();\n\n // Phone: common Nigerian / international patterns\n const phoneMatch = text.match(\n /(?:\\+?234[\\s-]?)?[0]\\d{9,10}\\b|\\+234[\\s-]?\\d{3}[\\s-]?\\d{3}[\\s-]?\\d{4}/,\n );\n\n const outcome = await ingestLead({\n name: nameRaw || undefined,\n gmail: gmail || undefined,\n phone: phoneMatch?.[0]?.replace(/[\\s-]/g, \"\") || undefined,\n message: text.slice(0, 2000),\n source: \"email-intake\",\n source_url: from.match(/<([^>]+)>/)?.[1] || e.from || \"\",\n });\n\n console.log(\n `email-intake <- ${gmail} => ${outcome.decision} (${\n (outcome as { reason?: string }).reason ?? \"\"\n })`,\n );\n}\n",[],[196,369],["D",1788495175441],[246],[196],"d1e73a5d-e4c4-4e91-9abe-00a0aa4b833c",["D",1788495223471],[],[34,50],["D",1788495223471],[244],[34],"4f07185f-72ef-41c9-90ff-d977099001bd","schema.sql","8257134c-a7b3-11f1-a292-1607ee4eb77e",48,["D",1788493785335],"-- =============================================================================\n-- DLX Designs — verified leads (Supabase / PostgreSQL schema)\n-- Run this in the Supabase SQL editor (Dashboard > SQL > New query) to create\n-- the table the val mirrors verified leads + pipeline state into. The val's\n-- native SQLite store works even without Supabase; setting SUPABASE_URL +\n-- SUPABASE_SERVICE_ROLE_KEY turns this table into a live mirror.\n-- =============================================================================\n\ncreate extension if not exists pgcrypto;\n\ncreate table if not exists public.dlx_verified_leads (\n id uuid primary key default gen_random_uuid(),\n name text not null,\n gmail text not null unique,\n phone text not null default '',\n score integer not null default 0 check (score between 0 and 100),\n source text not null default '',\n source_url text not null default '',\n evidence text not null default '',\n signal text not null default '',\n action text not null default '',\n status text not null default 'new',\n history jsonb not null default '[]'::jsonb,\n created_at timestamptz not null default now(),\n updated_at timestamptz not null default now()\n);\n\n-- Fast sorts: strongest opportunities first, then newest.\ncreate index if not exists dlx_verified_leads_score_idx\n on public.dlx_verified_leads(score desc);\ncreate index if not exists dlx_verified_leads_created_idx\n on public.dlx_verified_leads(created_at desc);\ncreate index if not exists dlx_verified_leads_status_idx\n on public.dlx_verified_leads(status);\ncreate index if not exists dlx_verified_leads_source_idx\n on public.dlx_verified_leads(source);\n\n-- Row Level Security: lock it down so only service-role (token) calls can\n-- read/write until you opt into per-user policies. Using the anon key from the\n-- browser WILL be blocked by default — intentional (this is internal data).\nalter table public.dlx_verified_leads enable row level security;\n",[],[205,386],["D",1788452485280],[242],[205],"697dc879-346a-476d-b8d5-6cb2817ec391","hunt.ts","a828a7f6-a813-11f1-8400-1607ee4eb77e",["D",1788493780394],"interval","// cron/hunt.ts\n// Scheduled hunt poller. Runs every interval; pulls genuine inquiries from the\n// configured real source (DLX_HUNT_URL) into the pipeline. Safe to run\n// repeatedly — duplicates are skipped by the unique email constraint.\nimport { hunt } from \"../backend/hunt.ts\";\n\nexport default async function (interval: Interval) {\n await hunt(interval.lastRunAt ?? new Date());\n}\n",[],[211,397],["D",1788493780394],[240],[211],"5a3d3f9e-c71d-491f-bc53-88b16a617083","pipeline.ts","7f43465c-a813-11f1-a948-1607ee4eb77e",34,["D",1788493711801],"// backend/pipeline.ts\n// The DLX Designs sales pipeline stages. Every verified lead moves through\n// these; each transition is recorded in the lead's history so the business can\n// learn which stages convert and from which sources.\nexport const STAGES = [\n \"new\",\n \"qualified\",\n \"contacted\",\n \"responded\",\n \"consultation\",\n \"quoted\",\n \"deposit_paid\",\n \"ordered\",\n \"won\",\n \"lost\",\n] as const;\n\nexport type Stage = (typeof STAGES)[number];\n\nexport function isStage(s: string): s is Stage {\n return (STAGES as readonly string[]).includes(s);\n}\n\nexport const STAGE_LABELS: Record = {\n new: \"New\",\n qualified: \"Qualified\",\n contacted: \"Contacted\",\n responded: \"Responded\",\n consultation: \"Consultation\",\n quoted: \"Quoted\",\n deposit_paid: \"Deposit Paid\",\n ordered: \"Ordered\",\n won: \"Won\",\n lost: \"Lost\",\n};\n\n/** \"Won\" and \"lost\" are terminal — nothing advances after them. */\nexport function nextStage(current: Stage): Stage | null {\n if (current === \"won\" || current === \"lost\") return null;\n const idx = STAGES.indexOf(current);\n return STAGES[idx + 1] ?? null;\n}\n",[],[217,408],"c410888c-4a87-4470-a3df-8bcc00cc17b4","a81c406a-a813-11f1-8400-1607ee4eb77e",54,["D",1788494704673],"// backend/hunt.ts\n// HUNT — the autonomous ingestion layer.\n//\n// This is the honest form of \"hunting\": it watches a REAL, configured source\n// and feeds genuinely captured inquiries through the qualifier into the\n// pipeline. It never invents people, names, or signals — it only ingests what\n// the source actually contains.\n//\n// Configure one or more real sources:\n// DLX_HUNT_URL — a JSON or CSV endpoint / published feed.\n// DLX_HUNT_TOKEN — optional bearer token sent as `Authorization: Bearer …`\n// (e.g. for a private feed).\n//\n// JSON response shape (either a bare array or wrapped):\n// { \"leads\": [ { name, gmail, phone, message, source, source_url } , ... ] }\n//\n// CSV: a published Google Sheet CSV is accepted directly. First row must be a\n// header that includes at least `gmail` (plus optional name, phone, message,\n// source, source_url). Google Sheets: File → Share → Publish to web → CSV,\n// copy the link, use it as DLX_HUNT_URL.\n//\n// Deduplication is automatic: the store's unique gmail constraint means a lead\n// already in the pipeline is skipped, so the poller is idempotent and safe to\n// run on a cron.\nimport { ingestLead, type InboundLead } from \"./worker/ingest.ts\";\n\nexport interface HuntResult {\n source: string;\n fetched: number;\n qualified: number;\n rejected: number;\n duplicates: number;\n}\n\nexport async function hunt(now = new Date()): Promise {\n const url = Deno.env.get(\"DLX_HUNT_URL\");\n const token = Deno.env.get(\"DLX_HUNT_TOKEN\");\n const result: HuntResult = { source: url ?? \"(none)\", fetched: 0, qualified: 0, rejected: 0, duplicates: 0 };\n\n if (!url) {\n console.log(`hunt: no DLX_HUNT_URL configured at ${now.toISOString()}`);\n return result;\n }\n\n const headers: Record = { Accept: \"text/csv, application/json, text/plain\" };\n if (token) headers.Authorization = `Bearer ${token}`;\n\n let text: string;\n try {\n const res = await fetch(url, { headers });\n if (!res.ok) throw new Error(`HTTP ${res.status}`);\n text = await res.text();\n } catch (err) {\n console.error(\"hunt: fetch failed from source:\", err);\n return result;\n }\n\n const trimmed = text.trim();\n const looksJson = trimmed.startsWith(\"{\") || trimmed.startsWith(\"[\");\n const rows: Record[] = looksJson\n ? jsonlize(trimmed)\n : csvlize(trimmed);\n\n for (const item of rows) {\n if (!item.gmail) continue; // evidence-less rows are ignored, never invented\n result.fetched++;\n const outcome = await ingestLead(item as unknown as InboundLead);\n if (outcome.decision === \"pass\") {\n result.qualified++;\n } else if (\n typeof outcome.reason === \"string\" &&\n outcome.reason.includes(\"duplicate\")\n ) {\n result.duplicates++;\n } else {\n result.rejected++;\n }\n }\n\n console.log(\n `hunt ${now.toISOString()} source=${result.source} fetched=${result.fetched} qualified=${result.qualified} rejected=${result.rejected} duplicates=${result.duplicates}`,\n );\n return result;\n}\n\nfunction jsonlize(text: string): Record[] {\n let payload: unknown;\n try {\n payload = JSON.parse(text);\n } catch {\n return [];\n }\n if (Array.isArray(payload)) return payload as Record[];\n if (Array.isArray((payload as { leads?: unknown[] })?.leads)) {\n return (payload as { leads: unknown[] }).leads as Record[];\n }\n return [];\n}\n\n/** Minimal RFC-4180-ish CSV parser (handles quoted fields and embedded commas). */\nfunction csvlize(text: string): Record[] {\n const lines = splitCsvLines(text);\n if (lines.length < 2) return [];\n const header = parseCsvLine(lines[0]).map((h) => h.trim().toLowerCase());\n const gidx = header.indexOf(\"gmail\");\n if (gidx === -1) return []; // must have a gmail column\n const out: Record[] = [];\n for (let i = 1; i < lines.length; i++) {\n const cells = parseCsvLine(lines[i]);\n if (!cells[gidx]?.trim()) continue;\n const row: Record = {};\n header.forEach((h, j) => {\n if (h) row[h] = (cells[j] ?? \"\").trim();\n });\n if (row.gmail) out.push(row);\n }\n return out;\n}\n\nfunction splitCsvLines(text: string): string[] {\n const out: string[] = [];\n let cur = \"\";\n let inQ = false;\n for (let i = 0; i < text.length; i++) {\n const ch = text[i];\n if (ch === '\"') {\n inQ = !inQ;\n cur += ch;\n } else if (ch === \"\\n\" && !inQ) {\n out.push(cur);\n cur = \"\";\n } else {\n cur += ch;\n }\n }\n if (cur.trim()) out.push(cur);\n return out;\n}\n\nfunction parseCsvLine(line: string): string[] {\n const cells: string[] = [];\n let cur = \"\";\n let inQ = false;\n for (let i = 0; i < line.length; i++) {\n const ch = line[i];\n if (ch === '\"') {\n if (inQ && line[i + 1] === '\"') {\n cur += '\"';\n i++;\n } else {\n inQ = !inQ;\n }\n } else if (ch === \",\" && !inQ) {\n cells.push(cur);\n cur = \"\";\n } else {\n cur += ch;\n }\n }\n cells.push(cur);\n return cells;\n}\n",[],[217,397],"3362ae21-a67d-4024-9bca-509f936d2314","qualifier.ts","8248f410-a7b3-11f1-a292-1607ee4eb77e","31184a6c-2262-49a1-9d1e-4a43608cddfb",36,["D",1788493720408],"// backend/worker/qualifier.ts\n// The strict qualification engine for DLX Designs.\n//\n// Two layers:\n// 1. Hard disqualifiers — instant reject on low-budget language.\n// 2. Scoring matrix (0-100) — high-tier buying signals raise the score.\n// A lead only passes when (a) no disqualifier fires AND (b) its score clears\n// the PASS_THRESHOLD. Inputs that fail are rejected and never persisted.\nimport { llmRefineScore } from \"./llm.ts\";\n\nexport interface InboundLead {\n name?: unknown;\n gmail?: unknown;\n phone?: unknown;\n message?: unknown;\n source?: unknown;\n source_url?: unknown;\n evidence?: unknown;\n}\n\nexport interface QualifiedLead {\n name: string;\n gmail: string;\n phone: string;\n score: string; // e.g. \"80/100%\"\n source: string;\n source_url: string;\n evidence: string;\n signal: string;\n action: string; // recommended next action\n}\n\nexport interface QualifyOutcome {\n decision: \"pass\" | \"reject\";\n lead?: QualifiedLead;\n reason: string;\n numericScore?: number;\n}\n\nconst DISQUALIFIERS = [\n \"cheap\",\n \"affordable\",\n \"under 20k\",\n \"under20k\",\n \"20k\",\n \"discount\",\n \"thrift\",\n \"okirika\",\n];\n\n// High-tier signals -> points toward intent. Matched words are deduped so a\n// single mention counts once.\nconst HIGH_TIER: Record = {\n bespoke: 18,\n \"3-piece\": 12,\n \"three-piece\": 12,\n \"3 piece\": 12,\n cashmere: 15,\n groom: 20,\n chieftaincy: 20,\n executive: 14,\n wedding: 14,\n suit: 8,\n agbada: 8,\n quality: 5,\n premium: 8,\n lagos: 6,\n abeokuta: 6,\n lekki: 10,\n \"victoria island\": 10,\n nigeria: 3,\n};\n\nconst BASE_SCORE = 25;\nconst PASS_THRESHOLD = 50;\nconst MAX_SCORE = 100;\n\nconst EMAIL_RE = /[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}/;\n\n/**\n * Strict, self-contained heuristic qualification. Never throws; never\n * fabricates data. Works with or without an LLM key.\n */\nexport async function qualifyLead(inbound: InboundLead): Promise {\n const message = (String(inbound?.message ?? \"\")).toLowerCase();\n const gmailRaw = String(inbound?.gmail ?? \"\").trim();\n const source = String(inbound?.source ?? (inbound?.gmail ? \"inbound-form\" : \"\")).trim();\n\n // --- Layer 1: hard disqualifiers -------------------------------------\n for (const d of DISQUALIFIERS) {\n if (message.includes(d)) {\n return {\n decision: \"reject\",\n reason: `disqualifier '${d}' present in message (low-budget signal)`,\n };\n }\n }\n\n // --- Validation -------------------------------------------------------\n const gmailMatch = gmailRaw.toLowerCase().match(EMAIL_RE);\n if (!gmailMatch) {\n return { decision: \"reject\", reason: \"no valid email address supplied\" };\n }\n const gmail = gmailMatch[0].toLowerCase();\n\n const name = (String(inbound?.name ?? \"\").trim() ||\n gmail.split(\"@\")[0]).slice(0, 80);\n if (!name) {\n return { decision: \"reject\", reason: \"no name supplied\" };\n }\n\n // --- Layer 2: scoring matrix ------------------------------------------\n let score = BASE_SCORE;\n const matched: string[] = [];\n for (const [kw, pts] of Object.entries(HIGH_TIER)) {\n if (message.includes(kw)) {\n score += pts;\n matched.push(kw);\n }\n }\n score = Math.min(MAX_SCORE, score);\n const numericScore = score;\n\n if (numericScore < PASS_THRESHOLD) {\n return {\n decision: \"reject\",\n reason:\n `intent score ${numericScore}/100 below ${PASS_THRESHOLD} threshold ` +\n `(insufficient buying signal — no bespoke/luxury context detected)`,\n numericScore,\n };\n }\n\n // Build a human-readable signal from the strongest matched keywords.\n const signal =\n matched.length > 0\n ? `Interest detected: ${matched.slice(0, 4).join(\", \")}`\n : \"High-intent luxury menswear inquiry\";\n\n // --- Layer 3: optional LLM refinement --------------------------------\n let finalScore = numericScore;\n let finalSignal = signal;\n const refined = await llmRefineScore(inbound, numericScore);\n if (refined) {\n finalScore = refined.numericScore;\n if (refined.signal) finalSignal = refined.signal;\n }\n\n return {\n decision: \"pass\",\n lead: {\n name,\n gmail,\n phone: String(inbound?.phone ?? \"\").trim().slice(0, 40),\n score: `${finalScore}/100%`,\n source: source || \"inbound-form\",\n source_url: String(inbound?.source_url ?? \"\").trim().slice(0, 500),\n evidence: (String(inbound?.evidence ?? \"\") ||\n String(inbound?.message ?? \"\")).trim().slice(0, 2000),\n signal: finalSignal,\n action: recommendAction(finalScore, finalSignal),\n },\n reason: \"verified high-intent lead\",\n numericScore: finalScore,\n };\n}\n\n/** Recommend the next sales action from the proven intent signal — never invented. */\nfunction recommendAction(score: number, signal: string): string {\n const s = signal.toLowerCase();\n if (/(groom|chieftaincy|wedding|consultation|quote)/.test(s)) {\n return \"Book consultation — upcoming occasion, quote at first meeting\";\n }\n if (score >= 80) {\n return \"High priority — contact within 24h to arrange consultation\";\n }\n if (score >= 65) {\n return \"Contact to confirm budget & timeline, then quote\";\n }\n return \"Nurture — follow up to confirm occasion & budget before quoting\";\n}\n",[],[217,431,423],"worker","bab5c45c-d890-4bf3-9ff4-227b1782d3e6","llm.ts","735a5200-a7b3-11f1-8441-1607ee4eb77e",27,["D",1788493422605],"// backend/worker/llm.ts\n// PLUGGABLE, OPTIONAL AI score-refinement adapter.\n//\n// The engine is DETERMINISTIC-FIRST and fully production-ready WITHOUT any AI\n// key: the strict heuristic scoring matrix in qualifier.ts is authoritative\n// and never depends on this module. That removes OpenAI (or any provider) as a\n// single point of failure.\n//\n// When a key IS present, the matching provider refines the score and writes a\n// natural buying signal. Providers are tried in order, free tiers first:\n// 1. GEMINI_API_KEY — Google AI Studio free tier (no billing required)\n// 2. GROQ_API_KEY — Groq free tier (fast, generous quota)\n// 3. OPENAI_API_KEY — OpenAI (requires active billing)\n// If a provider fails (wrong key, quota, billing, network), we simply fall\n// through to the next one and ultimately to the heuristic — never an error.\nexport interface InboundLead {\n name?: unknown;\n gmail?: unknown;\n message?: unknown;\n source?: unknown;\n}\n\nexport async function llmRefineScore(\n inbound: InboundLead,\n heuristicScore: number,\n): Promise<{ numericScore: number; signal: string } | null> {\n const prompt = buildPrompt(inbound, heuristicScore);\n\n // 1) Gemini (free tier — recommended)\n const geminiKey = Deno.env.get(\"GEMINI_API_KEY\");\n if (geminiKey) {\n const r = await tryGemini(geminiKey, prompt, heuristicScore);\n if (r) return r;\n }\n\n // 2) Groq (free tier)\n const groqKey = Deno.env.get(\"GROQ_API_KEY\");\n if (groqKey) {\n const r = await tryGroq(groqKey, prompt, heuristicScore);\n if (r) return r;\n }\n\n // 3) OpenAI (only reached if OpenAI is provided and Gemini/Groq fail)\n const openaiKey = Deno.env.get(\"OPENAI_API_KEY\");\n if (openaiKey) {\n const r = await tryOpenAI(openaiKey, prompt, heuristicScore);\n if (r) return r;\n }\n\n return null; // caller falls back to the pure heuristic result\n}\n\nfunction buildPrompt(inbound: InboundLead, heuristicScore: number): string {\n return (\n `You are a lead qualification analyst for DLX Designs, a luxury bespoke ` +\n `traditional menswear brand for high-net-worth clients, grooms and corporate ` +\n `executives in Lagos and Abeokuta, Nigeria. ` +\n `A lead submitted: name=\"${String(inbound.name ?? \"\")}\", ` +\n `email=\"${String(inbound.gmail ?? \"\")}\", message=\"${String(inbound.message ?? \"\")}\". ` +\n `Return ONLY JSON: {\"score\":0-100,\"signal\":\"one-line buying intent\"}. ` +\n `Reject low budget demand (cheap, affordable, discount, thrift, okirika). ` +\n `The strict heuristic already scored this lead ${heuristicScore}/100.`\n );\n}\n\nasync function tryGemini(\n key: string,\n prompt: string,\n fallback: number,\n): Promise<{ numericScore: number; signal: string } | null> {\n try {\n const res = await fetch(\n \"https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-flash:generateContent?key=\" +\n key,\n {\n method: \"POST\",\n headers: { \"Content-Type\": \"application/json\" },\n body: JSON.stringify({\n contents: [{ parts: [{ text: prompt + \" Respond with JSON only.\" }] }],\n }),\n },\n );\n if (!res.ok) throw new Error(`Gemini ${res.status}`);\n const data = await res.json();\n const text = data?.candidates?.[0]?.content?.parts?.[0]?.text ?? \"{}\";\n const cleaned = text.replace(/```json|```/g, \"\").trim();\n return normalize(JSON.parse(cleaned), fallback);\n } catch (err) {\n console.error(\"Gemini refine failed:\", err);\n return null;\n }\n}\n\nasync function tryGroq(\n key: string,\n prompt: string,\n fallback: number,\n): Promise<{ numericScore: number; signal: string } | null> {\n try {\n const res = await fetch(\"https://api.groq.com/openai/v1/chat/completions\", {\n method: \"POST\",\n headers: {\n \"Content-Type\": \"application/json\",\n Authorization: `Bearer ${key}`,\n },\n body: JSON.stringify({\n model: \"llama-3.3-70b-versatile\",\n temperature: 0,\n response_format: { type: \"json_object\" },\n messages: [\n { role: \"system\", content: \"You output strict JSON only.\" },\n { role: \"user\", content: prompt },\n ],\n }),\n });\n if (!res.ok) throw new Error(`Groq ${res.status} ${await res.text()}`);\n const data = await res.json();\n const parsed = JSON.parse(data.choices?.[0]?.message?.content ?? \"{}\");\n return normalize(parsed, fallback);\n } catch (err) {\n console.error(\"Groq refine failed:\", err);\n return null;\n }\n}\n\nasync function tryOpenAI(\n key: string,\n prompt: string,\n fallback: number,\n): Promise<{ numericScore: number; signal: string } | null> {\n try {\n const res = await fetch(\"https://api.openai.com/v1/chat/completions\", {\n method: \"POST\",\n headers: {\n \"Content-Type\": \"application/json\",\n Authorization: `Bearer ${key}`,\n },\n body: JSON.stringify({\n model: \"gpt-4o-mini\",\n temperature: 0,\n response_format: { type: \"json_object\" },\n messages: [\n { role: \"system\", content: \"You output strict JSON only.\" },\n { role: \"user\", content: prompt },\n ],\n }),\n });\n if (!res.ok) throw new Error(`OpenAI ${res.status}`);\n const data = await res.json();\n const parsed = JSON.parse(data.choices?.[0]?.message?.content ?? \"{}\");\n return normalize(parsed, fallback);\n } catch (err) {\n console.error(\"OpenAI refine failed:\", err);\n return null;\n }\n}\n\nfunction normalize(\n parsed: { score?: unknown; signal?: unknown },\n fallbackScore: number,\n): { numericScore: number; signal: string } {\n const s = Number(parsed?.score);\n // Blend heuristic and LLM, keeping the heuristic as a floor.\n const numericScore = Number.isFinite(s)\n ? Math.max(fallbackScore, Math.min(100, Math.round(s)))\n : fallbackScore;\n const signal =\n typeof parsed?.signal === \"string\" ? parsed.signal.slice(0, 200) : \"\";\n return { numericScore, signal };\n}\n",[],[217,431,433],"6f49607f-22c7-466b-b2a1-44928124c7e4","ingest.ts","824ff7c4-a7b3-11f1-a292-1607ee4eb77e",37,["D",1788493725323],"// backend/worker/ingest.ts\n// Ingest worker — the bridge between real inbound traffic and the verified\n// store. Runs the strict qualifier; only passing, high-intent leads persist.\n// Rejects are returned (and never stored).\nimport { insertLead } from \"../../lib/store.ts\";\nimport { qualifyLead, type InboundLead } from \"./qualifier.ts\";\n\nexport async function ingestLead(inbound: InboundLead) {\n const outcome = await qualifyLead(inbound);\n if (outcome.decision === \"reject\") {\n return outcome;\n }\n\n const lead = outcome.lead!;\n const scoreInt = Math.min(100, Math.max(0, outcome.numericScore ?? 0));\n\n const inserted = await insertLead({\n name: lead.name,\n gmail: lead.gmail,\n phone: lead.phone,\n score: scoreInt,\n source: lead.source,\n source_url: lead.source_url,\n evidence: lead.evidence,\n signal: lead.signal,\n action: lead.action,\n status: \"new\",\n });\n\n if (!inserted) {\n return {\n decision: \"reject\",\n reason: \"duplicate email — lead already verified\",\n numericScore: scoreInt,\n };\n }\n\n return {\n decision: \"pass\",\n lead: {\n id: inserted.id,\n name: inserted.name,\n gmail: inserted.gmail,\n phone: inserted.phone,\n score: `${inserted.score}/100%`,\n source: inserted.source,\n source_url: inserted.source_url,\n evidence: inserted.evidence,\n signal: inserted.signal,\n action: inserted.action,\n status: inserted.status,\n created_at: inserted.created_at,\n },\n reason: \"verified high-intent lead stored\",\n numericScore: inserted.score,\n };\n}\n",[],[217,431,441],"f46f34f4-1677-4619-a22e-94d613250e13",["D",1788452460127],[234,235,236],[217,431],["D",1788452460127],[233,237,238],[217],["D",1788452396724],["D",1789494036282],{"_110":109,"_117":-5},{"_110":109,"_111":112,"_113":-5,"_19":-5,"_114":115,"_116":31},{"_13":99,"_38":100,"_53":46,"_101":54,"_102":-5,"_103":-5,"_104":-5,"_105":-5,"_106":-5,"_107":55,"_108":109},"ctx",{"_92":-5,"_137":493,"_494":495,"_496":497,"_498":31,"_499":31,"_500":31,"_501":69,"_502":-5,"_503":-7,"_504":-7,"_505":73,"_506":31,"_507":31,"_508":-5,"_509":31,"_510":31,"_511":31,"_512":-5},"ENV",{"_464":465,"_466":467,"_468":100,"_469":470,"_471":472,"_473":474,"_475":476,"_477":478,"_479":480,"_481":482,"_483":484,"_485":486,"_487":488,"_489":490,"_491":492},"REACT_APP_SENTRY_DSN","https://586020d86f05489db63a554690acf1e9@o1398359.ingest.sentry.io/6724499","RENDER_GIT_COMMIT","e87eb04e707cf64e7e532c8e8ec395f209e17f15","RENDER_GIT_BRANCH","POSTHOG_API_TOKEN","phc_nmoyuAkStEZuUp69rYlg1DZgrIt19Gx0Y4Hwf8tGgJP","POSTHOG_API_HOST","https://wart.val.town","WEB_API_TEMPLATE","https://{val}.web.val.run","APP_URL","https://www.val.town","MODULE_URL","https://esm.town","VANITY_WEB_API_TEMPLATE","https://{label}.val.run/","DENOLS_URL","https://lsp.val.town","NODE_ENV","production","CLOUDFLARE_TURNSTILE_SITE_KEY","0x4AAAAAACtaIZJH7s9rBMZ6","GOOGLE_CLIENT_ID","875989696899-jtsqkhlbdv5jrlsk8631p7ipjg3c6o62.apps.googleusercontent.com","STRIPE_PUBLISHABLE_KEY","pk_live_51N2zvWEpZTTLipIkVEwXrUG9AT0yeOOEYxszarZT1jpsZV58Ly6wlg3SltZpHN26CUalOBmUSY7adQRQiSkCZz2500qSXNkf8Y","API_URL","https://api.val.town",[],"preferencesKeybindings","default","preferencesTheme","light","preferencesFixers","preferencesCopilot","preferencesSelectionAi","preferencesCloseBrackets","preferencesChatPanelState","preferencesDprint","preferencesDenofmt","notificationCount","flagDenols","isBusinessOrgMember","businessOrgHandle","isBusinessOrgAdmin","isHigherTownieTier","navCollapsed","currentOrgId"]

Diff · v64