[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10},"actionData","errors","root",{"_191":192,"_193":194},"routes/_app.x.$handle.$name",{"_96":95,"_32":44,"_104":86,"_27":105,"_13":14,"_106":-5,"_11":107,"_108":109,"_110":111,"_112":113,"_114":115,"_20":-5,"_116":-5,"_117":-5,"_118":25,"_119":120,"_121":58,"_122":-5,"_123":-5,"_124":125},"routes/_app.x.$handle.$name.branch.$branch.version.$version.diff",{"_11":12,"_13":14,"_15":16,"_17":-5,"_18":-5},"project",{"_42":43,"_32":44,"_37":45,"_46":-5,"_47":25,"_48":49,"_50":49,"_51":52,"_53":-5,"_18":54,"_55":56,"_57":58,"_59":-5,"_60":-5,"_61":62,"_63":64,"_65":-5,"_66":-5,"_67":-5,"_68":-5,"_69":70,"_71":70,"_72":73,"_74":73,"_75":73,"_76":70,"_77":70,"_78":79,"_80":81,"_82":83,"_84":25},"version",6,"changes",[19],"aiChatInfo","userId",{"_20":21,"_22":23,"_24":25,"_26":25,"_27":28,"_29":28},"file",{"_30":31,"_32":28,"_33":20,"_34":36,"_37":38,"_39":14,"_40":41,"_24":25},"previous",{"_30":31,"_32":28,"_33":20,"_34":35},"deleted",false,"deletedHere","path","index.html","latestPath","fileId","acee3b16-852c-11f1-a666-1607ee4eb77e","name","fileType","content","\n\n\n\n \n \n Mitch's Multifarious Three-Headed Hydra Chatbot\n \n\n\n\n

Mitch's Multifarious Three-Headed Hydra Chatbot

\n Ask a question. Every answer is written one word at a time. Three models take turns, and\n each one just reads everything written so far before adding the next word. You can see\n which model produced each word from the pretty colors.\n

\n Many people assume that LLMs have something akin to consciousness that persists during\n each session, but this experiment shows that that assumption is unnecessary. In reality,\n LLMs have separate thought processes for each individual token (i.e. a fragment of a word)\n that they produce.\n

\n\n

\n \n Claude\n (anthropic/claude-haiku-4.5)\n \n \n Gemini\n (google/gemini-2.5-flash)\n \n \n Llama\n (meta-llama/llama-3.1-8b-instruct)\n \n

\n\n \n\n

\n Ask a question below to watch the three heads write an answer together.\n

\n\n \n

\n\n \n\n \n\n\n\n","\n\n\n\n \n \n Mitch's Multifarious Three-Headed Hydra Chatbot\n \n\n\n\n

Mitch's Multifarious Three-Headed Hydra Chatbot

\n Ask a question. Every answer is written one word at a time. Three models take turns, and\n each one just reads everything written so far before adding the next word. They have\n not\n been told they are playing a turn-taking game; this is just how they work. You can see\n which model produced each word from the pretty colors.\n

\n\n

\n \n Claude\n (anthropic/claude-haiku-4.5)\n \n \n Gemini\n (google/gemini-2.5-flash)\n \n \n Llama\n (meta-llama/llama-3.1-8b-instruct)\n \n

\n\n \n\n

\n Ask a question below to watch the three heads write an answer together.\n

\n\n \n

\n\n \n\n \n\n\n\n","createdAt",["D",1784656413762],"branchVersion","branchId","c0965a46-0248-4c1a-8e1a-6da8fefe1c72","id","9fc1fa11-5119-48bd-834a-aa5ae54463e8","three-headed-hydra",["D",1784656158273],"deletedAt","legacyEnvVars","privacy","public","httpPrivacy","description","LLM next-word demo: Claude, Gemini & Llama take turns","imageUrl","c878cd0e-42f2-4d5f-b236-a8bb922062ca","pinnedFileId","7f21020e-852c-11f1-9143-1607ee4eb77e","pinPreview",true,"pinMigrated","screenshotVersion","likeCount",1,"orgId","805def35-af33-4bd3-b687-47a0923f59a5","tursoName","tursoHostname","tursoMeasuredAt","pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_96":95,"_103":99},"org",{"_96":95,"_97":-5,"_98":99,"_18":54,"_100":101,"_102":25},"branch",{"_13":85,"_32":86,"_42":41,"_87":43,"_88":-5,"_89":-5,"_90":-5,"_91":-5,"_92":-5,"_93":44,"_94":95},"isOwned",14,"main","projectId","forkedBranchId","forkedBranchVersion","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","tobeofuse","handle","displayName","avatar","https://img.clerk.com/eyJ0eXBlIjoicHJveHkiLCJzcmMiOiJodHRwczovL2ltYWdlcy5jbGVyay5kZXYvb2F1dGhfZ2l0aHViL2ltZ18yamFZTkZwTnNHSHJUbTFpclcwUDZzMXlCUEcifQ","tier","free","canRestrictVals","profileImageUrl","branchName","","adoptOrgId",{"_42":43,"_32":44,"_37":187,"_46":-5,"_47":25,"_48":49,"_50":49,"_51":52,"_53":-5,"_18":54,"_55":56,"_57":58,"_59":-5,"_60":-5,"_61":62,"_63":64,"_65":-5,"_66":-5,"_67":-5,"_68":-5,"_69":70,"_71":70,"_72":73,"_74":73,"_75":73,"_76":70,"_77":70,"_78":188,"_80":189,"_82":190,"_84":25},"tree",[170,171,172,173],"files",[170,171,172,173],"counts",{"_160":137,"_161":62,"_162":137,"_163":164,"_165":166},"contents",[126,127,128,129],"upstreamProject","upstreamBranch","emptyEnvVars","flags",[],"hasHttpFiles","grants","orgs","customDomains",{},{"_42":152,"_32":153,"_30":154,"_133":-5,"_40":41,"_39":155,"_135":58,"_37":156,"_13":137,"_33":157,"_34":158,"_24":25,"_139":-5,"_27":159},{"_42":149,"_32":28,"_30":31,"_133":-5,"_40":41,"_39":14,"_135":25,"_37":150,"_13":62,"_33":20,"_34":36,"_24":25,"_139":-5,"_27":151},{"_42":141,"_32":142,"_30":56,"_133":-5,"_40":41,"_39":143,"_135":58,"_37":144,"_13":62,"_33":145,"_34":146,"_24":25,"_139":147,"_27":148},{"_42":130,"_32":131,"_30":132,"_133":-5,"_40":41,"_39":134,"_135":58,"_37":136,"_13":137,"_33":20,"_34":138,"_24":25,"_139":-5,"_27":140},"0bd26beb-f5d6-4b5b-b691-dc0ad06a8c55","styles.css","ae29af88-852c-11f1-81e1-1607ee4eb77e","parentId",4,"latest",["D",1784656237214],0,":root {\n --claude: #c15f3c;\n --claude-bg: #f6e5de;\n --gemini: #3b73e0;\n --gemini-bg: #dde8fb;\n --llama: #7a5cd0;\n --llama-bg: #e8e1f7;\n\n --ink: #1f2328;\n --muted: #6b7280;\n --line: #e2e2e2;\n --panel: #ffffff;\n --page: #f4f3ef;\n}\n\n* {\n box-sizing: border-box;\n}\n\nbody {\n margin: 0;\n background: var(--page);\n color: var(--ink);\n font-family: \"Segoe UI\", system-ui, -apple-system, Helvetica, Arial, sans-serif;\n line-height: 1.55;\n}\n\n.app {\n max-width: 760px;\n margin: 0 auto;\n padding: 24px 20px 32px;\n min-height: 100vh;\n display: flex;\n flex-direction: column;\n}\n\n.header h1 {\n font-size: 1.55rem;\n margin: 0 0 6px;\n}\n\n.subhead {\n font-size: 1.02rem;\n font-weight: 600;\n color: var(--ink);\n margin: 0 0 10px;\n}\n\n.tagline {\n margin: 8px 0;\n color: var(--ink);\n font-size: 0.92rem;\n}\n\n/* Legend ------------------------------------------------------------------ */\n.legend {\n display: flex;\n flex-wrap: wrap;\n gap: 16px;\n margin: 18px 0 10px;\n padding: 10px 14px;\n background: var(--panel);\n border: 1px solid var(--line);\n border-radius: 8px;\n font-size: 0.88rem;\n}\n\n.legend-item {\n display: inline-flex;\n align-items: center;\n gap: 7px;\n}\n\n.legend-model {\n color: var(--muted);\n font-size: 0.78rem;\n}\n\n.swatch {\n width: 14px;\n height: 14px;\n border-radius: 3px;\n display: inline-block;\n}\n\n.head-claude {\n background: var(--claude);\n}\n\n.head-gemini {\n background: var(--gemini);\n}\n\n.head-llama {\n background: var(--llama);\n}\n\n/* Mode banner ------------------------------------------------------------- */\n.banner {\n margin: 6px 0 0;\n padding: 9px 12px;\n border-radius: 8px;\n font-size: 0.85rem;\n border: 1px solid var(--line);\n background: #fffaf0;\n color: #7a5b17;\n}\n\n/* Transcript -------------------------------------------------------------- */\n.transcript {\n flex: 1;\n margin: 16px 0;\n padding: 4px 2px;\n overflow-y: auto;\n}\n\n.empty-state {\n color: var(--muted);\n text-align: center;\n padding: 40px 10px;\n font-size: 0.95rem;\n}\n\n.message {\n margin: 0 0 18px;\n}\n\n.message .role {\n font-size: 0.72rem;\n letter-spacing: 0.05em;\n text-transform: uppercase;\n color: var(--muted);\n margin-bottom: 5px;\n}\n\n.bubble {\n border: 1px solid var(--line);\n border-radius: 10px;\n padding: 11px 14px;\n background: var(--panel);\n white-space: pre-wrap;\n word-wrap: break-word;\n}\n\n.message.user .bubble {\n background: #eef1f4;\n}\n\n/* Coloured words ---------------------------------------------------------- */\n.token {\n border-radius: 2px;\n padding: 1px 0;\n transition: background 0.15s ease;\n}\n\n.token.head-claude {\n background: var(--claude-bg);\n box-shadow: inset 0 -2px 0 var(--claude);\n}\n\n.token.head-gemini {\n background: var(--gemini-bg);\n box-shadow: inset 0 -2px 0 var(--gemini);\n}\n\n.token.head-llama {\n background: var(--llama-bg);\n box-shadow: inset 0 -2px 0 var(--llama);\n}\n\n.cursor {\n display: inline-block;\n width: 7px;\n height: 1.05em;\n vertical-align: text-bottom;\n background: var(--ink);\n margin-left: 1px;\n animation: blink 1s steps(2, start) infinite;\n}\n\n@keyframes blink {\n 50% {\n opacity: 0;\n }\n}\n\n/* Composer ---------------------------------------------------------------- */\n.composer {\n display: flex;\n gap: 8px;\n position: sticky;\n bottom: 0;\n padding-top: 6px;\n background: var(--page);\n}\n\n.prompt-input {\n flex: 1;\n padding: 11px 13px;\n border: 1px solid var(--line);\n border-radius: 8px;\n font-size: 0.98rem;\n background: var(--panel);\n color: var(--ink);\n}\n\n.prompt-input:focus {\n outline: 2px solid #c9d3e4;\n border-color: #c9d3e4;\n}\n\n.send-button {\n padding: 0 20px;\n border: none;\n border-radius: 8px;\n background: var(--ink);\n color: #fff;\n font-size: 0.95rem;\n cursor: pointer;\n}\n\n.send-button:disabled {\n opacity: 0.5;\n cursor: default;\n}\n\n.error {\n color: #b42318;\n font-size: 0.85rem;\n margin-top: 6px;\n}\n\n/* Caveats + source link (footer below the app) ---------------------------- */\n.caveats {\n max-width: 760px;\n margin: 0 auto;\n padding: 22px 20px 40px;\n border-top: 1px solid var(--line);\n color: var(--muted);\n font-size: 0.85rem;\n line-height: 1.55;\n}\n\n.caveats h2 {\n margin: 0 0 10px;\n font-size: 0.72rem;\n font-weight: 600;\n letter-spacing: 0.06em;\n text-transform: uppercase;\n color: var(--ink);\n}\n\n.caveats ul {\n margin: 0;\n padding-left: 18px;\n}\n\n.caveats li {\n margin: 0 0 9px;\n}\n\n.caveats li:last-child {\n margin-bottom: 0;\n}\n\n.source-link {\n margin: 16px 0 0;\n}\n\n.source-link a {\n color: var(--gemini);\n text-decoration: none;\n}\n\n.source-link a:hover {\n text-decoration: underline;\n}\n","label",[131],"2207486f-2344-4c59-816e-dcb00afde73b","main.tsx",2,["D",1784656228395],"http","/**\n * Mitch's Multifarious Three-Headed Hydra Chatbot -- Val Town edition.\n *\n * A single Deno HTTP val (no Express). A response is grown one word at a time,\n * with three models taking turns: Claude -> Gemini -> Llama -> Claude -> ...\n * After every word the whole answer-so-far is handed to the next head (as an\n * assistant \"prefill\"), which predicts the next word. Words stream to the\n * browser over Server-Sent Events, colour-coded by which model produced them.\n *\n * `staticHTTPServer` serves the frontend files (index.html, styles.css, app.js),\n * serves index.html at \"/\", and exposes a \"/source\" redirect to this val's code.\n */\nimport { staticHTTPServer } from \"https://esm.town/v/std/utils/index.ts\";\n\nconst OPENROUTER_URL = \"https://openrouter.ai/api/v1/chat/completions\";\n\n// Cheap models, one small model per family. Overridable via env vars.\nconst MODELS: Record = {\n claude: Deno.env.get(\"HYDRA_CLAUDE_MODEL\") || \"anthropic/claude-haiku-4.5\",\n gemini: Deno.env.get(\"HYDRA_GEMINI_MODEL\") || \"google/gemini-2.5-flash\",\n llama: Deno.env.get(\"HYDRA_LLAMA_MODEL\") || \"meta-llama/llama-3.1-8b-instruct\",\n};\nconst HEADS = [\"claude\", \"gemini\", \"llama\"];\n\n// Request a few tokens per turn but keep only the first whole word, so we avoid\n// mid-word tokenization artifacts.\nconst WORD_TOKEN_BUDGET = 8;\n\n// Appended to the prompt so the heads write one flowing paragraph of plain prose\n// instead of Markdown documents (whose headings and bullets interrupt sentences).\nconst STYLE_HINT =\n \"\\n\\nAnswer in one short paragraph of plain, conversational prose. Do not use \" +\n \"Markdown, headings, bullet points, numbered lists, or bold text.\";\n\n// --- Demo-mode simulated response (used when no OPENROUTER_API_KEY is set) ---\nconst SCRIPT =\n \"That is a great question. Language models do not write a whole sentence at \" +\n \"once. Instead, they predict the next word from everything written so far. \" +\n \"Here, three different models take turns, yet the result still reads as a \" +\n \"single coherent thought. That is the whole trick behind modern text generation.\";\nconst MOCK_WORDS = SCRIPT.split(\" \").map((word, i) => (i === 0 ? word : \" \" + word));\n\nfunction mockNextToken(index: number) {\n if (index >= MOCK_WORDS.length) return { token: \"\", done: true };\n return { token: MOCK_WORDS[index], done: false };\n}\n\nfunction isLive() {\n return Boolean(Deno.env.get(\"OPENROUTER_API_KEY\"));\n}\n\nfunction getStatus() {\n return { mode: isLive() ? \"live\" : \"mock\", models: MODELS };\n}\n\nconst sleep = (ms: number) => new Promise((r) => setTimeout(r, ms));\n\nasync function openRouterNextToken(head: string, prompt: string, accumulated: string) {\n const model = MODELS[head];\n const controller = new AbortController();\n const timeout = setTimeout(() => controller.abort(), 20000);\n\n try {\n const res = await fetch(OPENROUTER_URL, {\n method: \"POST\",\n signal: controller.signal,\n headers: {\n Authorization: `Bearer ${Deno.env.get(\"OPENROUTER_API_KEY\")}`,\n \"Content-Type\": \"application/json\",\n \"HTTP-Referer\": \"https://val.town\",\n \"X-Title\": \"Mitch's Multifarious Three-Headed Hydra Chatbot\",\n },\n // Send the prompt (plus a style hint), then the answer so far as an\n // assistant \"prefill\", and ask it to continue. Continuing (not restarting)\n // is why each word arrives with its own natural spacing -- and why the\n // heads must support prefill (OpenAI models don't, hence no GPT head).\n body: JSON.stringify({\n model,\n max_tokens: WORD_TOKEN_BUDGET,\n messages: accumulated\n ? [\n { role: \"user\", content: prompt + STYLE_HINT },\n { role: \"assistant\", content: accumulated },\n ]\n : [{ role: \"user\", content: prompt + STYLE_HINT }],\n }),\n });\n\n if (!res.ok) {\n const detail = await res.text().catch(() => \"\");\n throw new Error(`OpenRouter ${res.status}: ${detail.slice(0, 200)}`);\n }\n\n const data = await res.json();\n const continuation = data?.choices?.[0]?.message?.content ?? \"\";\n // Keep the first whole word, preserving any leading whitespace the model\n // produced (that space -- or its absence -- tells us whether this is a new\n // word or a continuation of the previous one).\n const word = (continuation.match(/^\\s*\\S+/) || [\"\"])[0];\n return { token: word, done: word === \"\" };\n } finally {\n clearTimeout(timeout);\n }\n}\n\nasync function getNextToken(\n { mode, head, prompt, accumulated, index }: {\n mode: string;\n head: string;\n prompt: string;\n accumulated: string;\n index: number;\n },\n) {\n if (mode === \"mock\") return mockNextToken(index);\n return openRouterNextToken(head, prompt, accumulated);\n}\n\nfunction handleGenerate(req: Request): Response {\n const url = new URL(req.url);\n const prompt = (url.searchParams.get(\"prompt\") || \"\").slice(0, 2000).trim();\n const max = Math.min(parseInt(url.searchParams.get(\"max\") || \"200\", 10) || 200, 200);\n const mode = isLive() ? \"live\" : \"mock\";\n\n const encoder = new TextEncoder();\n let cancelled = false;\n\n const stream = new ReadableStream({\n async start(controller) {\n const send = (event: string, data: unknown) => {\n if (!cancelled) {\n controller.enqueue(encoder.encode(`event: ${event}\\ndata: ${JSON.stringify(data)}\\n\\n`));\n }\n };\n\n if (!prompt) {\n send(\"error\", { message: \"Please enter a prompt.\" });\n controller.close();\n return;\n }\n\n send(\"meta\", { mode, heads: HEADS });\n\n let accumulated = \"\";\n let produced = 0;\n let emptyStreak = 0;\n\n for (let turn = 0; produced < max && !cancelled; turn++) {\n const head = HEADS[turn % HEADS.length];\n\n let result;\n try {\n result = await getNextToken({ mode, head, prompt, accumulated, index: produced });\n } catch (err) {\n send(\"error\", { message: `The ${head} head failed: ${(err as Error).message}` });\n break;\n }\n\n if (cancelled || result.done) break;\n\n const raw = result.token || \"\";\n if (raw === \"\") {\n if (++emptyStreak >= HEADS.length) break;\n if (mode === \"mock\") await sleep(45);\n continue;\n }\n emptyStreak = 0;\n\n // Decide the spacing in front of this word: keep the model's own leading\n // whitespace when it gave any (its signal for a new word, and preserves\n // paragraph breaks); otherwise glue it on directly, EXCEPT force a space\n // when the text so far ends in sentence punctuation (models drop that\n // space inconsistently, which produced \"sturdy,low\" / \"markings.They're\").\n let piece: string;\n if (/^\\s/.test(raw) || accumulated === \"\") {\n piece = raw;\n } else if (/[,.;:!?]$/.test(accumulated) && !/^[,.;:!?)\\]}%…'’”]/.test(raw)) {\n piece = \" \" + raw;\n } else {\n piece = raw;\n }\n accumulated += piece;\n produced += 1;\n\n send(\"token\", { head, text: piece, index: produced });\n\n if (mode === \"mock\") await sleep(55);\n }\n\n if (!cancelled) send(\"done\", { text: accumulated, count: produced });\n controller.close();\n },\n cancel() {\n cancelled = true;\n },\n });\n\n return new Response(stream, {\n headers: {\n \"Content-Type\": \"text/event-stream\",\n \"Cache-Control\": \"no-cache\",\n \"Connection\": \"keep-alive\",\n \"X-Accel-Buffering\": \"no\",\n },\n });\n}\n\nexport default staticHTTPServer(import.meta.url, (req: Request) => {\n const { pathname } = new URL(req.url);\n if (pathname === \"/api/status\") return Response.json(getStatus());\n if (pathname === \"/api/generate\") return handleGenerate(req);\n return null; // defer to the static server (index.html, styles.css, app.js, /source)\n});\n","mitchs-three-headed-hydra",[142],"f5f896ae-f460-4934-b8f2-1e885fec621f",["D",1784656413762],[28],"1f870eda-610c-4ba6-a305-e4c602df5af9","app.js","af03c952-852c-11f1-8306-1607ee4eb77e",5,["D",1784656238638],"script","// Frontend for the three-headed hydra. Opens an SSE stream and paints each\n// word as it arrives, colour-coded by the model that produced it.\n\nconst form = document.getElementById(\"chat-form\");\nconst input = document.getElementById(\"prompt-input\");\nconst sendButton = document.getElementById(\"send-button\");\nconst transcript = document.getElementById(\"transcript\");\nconst banner = document.getElementById(\"mode-banner\");\n\nconst HEAD_LABELS = { claude: \"Claude\", gemini: \"Gemini\", llama: \"Llama\" };\n\nlet streaming = false;\n\n// Show whether the server is using real models or the built-in simulation,\n// and fill the legend with the exact model names in use.\nasync function loadStatus() {\n try {\n const res = await fetch(\"/api/status\");\n const status = await res.json();\n\n for (const [head, model] of Object.entries(status.models || {})) {\n const el = document.querySelector(`.legend-model[data-head=\"${head}\"]`);\n if (el) el.textContent = `(${model})`;\n }\n\n // Only surface a banner when we're NOT live, to explain the simulation.\n if (status.mode !== \"live\") {\n banner.textContent =\n \"Demo mode: no OPENROUTER_API_KEY is set, so a simulated response plays back instead. The word-by-word, head-by-head mechanics are identical.\";\n banner.hidden = false;\n }\n } catch {\n /* status is a nicety; ignore failures */\n }\n}\n\nfunction clearEmptyState() {\n const empty = transcript.querySelector(\".empty-state\");\n if (empty) empty.remove();\n}\n\nfunction addMessage(role, roleLabel) {\n const wrap = document.createElement(\"div\");\n wrap.className = `message ${role}`;\n\n const label = document.createElement(\"div\");\n label.className = \"role\";\n label.textContent = roleLabel;\n\n const bubble = document.createElement(\"div\");\n bubble.className = \"bubble\";\n\n wrap.append(label, bubble);\n transcript.append(wrap);\n transcript.scrollTop = transcript.scrollHeight;\n return bubble;\n}\n\nfunction scrollToBottom() {\n transcript.scrollTop = transcript.scrollHeight;\n}\n\nfunction setStreaming(on) {\n streaming = on;\n sendButton.disabled = on;\n input.disabled = on;\n sendButton.textContent = on ? \"Writing…\" : \"Send\";\n}\n\nfunction submitPrompt(prompt) {\n clearEmptyState();\n addMessage(\"user\", \"You\").textContent = prompt;\n\n const bubble = addMessage(\"assistant\", \"The Hydra\");\n const cursor = document.createElement(\"span\");\n cursor.className = \"cursor\";\n bubble.append(cursor);\n\n setStreaming(true);\n\n const url = `/api/generate?prompt=${encodeURIComponent(prompt)}`;\n const source = new EventSource(url);\n\n const finish = () => {\n cursor.remove();\n source.close();\n setStreaming(false);\n input.focus();\n };\n\n source.addEventListener(\"token\", (event) => {\n const { head, text } = JSON.parse(event.data);\n const span = document.createElement(\"span\");\n span.className = `token head-${head}`;\n span.textContent = text;\n span.title = HEAD_LABELS[head] || head;\n bubble.insertBefore(span, cursor);\n scrollToBottom();\n });\n\n source.addEventListener(\"error\", (event) => {\n // Distinguish a server-sent \"error\" event from a dropped connection.\n let message = \"The connection was interrupted.\";\n if (event.data) {\n try {\n message = JSON.parse(event.data).message || message;\n } catch {\n /* keep default */\n }\n }\n const err = document.createElement(\"div\");\n err.className = \"error\";\n err.textContent = message;\n bubble.parentElement.append(err);\n finish();\n });\n\n source.addEventListener(\"done\", finish);\n}\n\nform.addEventListener(\"submit\", (event) => {\n event.preventDefault();\n if (streaming) return;\n const prompt = input.value.trim();\n if (!prompt) return;\n input.value = \"\";\n submitPrompt(prompt);\n});\n\nloadStatus();\n",[153],"pulls","branches","forks","envVars",{"_167":62,"_169":137},"connections",{"_167":137,"_168":137},"count","unfilledCount","emptyCount",{"_42":152,"_32":153,"_30":154,"_133":-5,"_40":41,"_39":155,"_135":58,"_37":184,"_13":137,"_33":157,"_34":158,"_24":25,"_139":-5,"_175":185,"_27":186},{"_42":149,"_32":28,"_30":31,"_133":-5,"_40":41,"_39":14,"_135":25,"_37":181,"_13":62,"_33":20,"_34":36,"_24":25,"_139":-5,"_175":182,"_27":183},{"_42":141,"_32":142,"_30":56,"_133":-5,"_40":41,"_39":143,"_135":58,"_37":178,"_13":62,"_33":145,"_34":146,"_24":25,"_139":147,"_175":179,"_27":180},{"_42":130,"_32":131,"_30":132,"_133":-5,"_40":41,"_39":134,"_135":58,"_37":174,"_13":137,"_33":20,"_34":138,"_24":25,"_139":-5,"_175":176,"_27":177},["D",1784656237214],"children",[],[131],["D",1784656228395],[],[142],["D",1784656413762],[],[28],["D",1784656238638],[],[153],["D",1784656158273],{"_96":95,"_103":99},{"_96":95,"_97":-5,"_98":99,"_18":54,"_100":101,"_102":25},{"_13":85,"_32":86,"_42":41,"_87":43,"_88":-5,"_89":-5,"_90":-5,"_91":-5,"_92":-5,"_93":44,"_94":95},"ctx",{"_78":-5,"_123":224,"_225":226,"_227":228,"_229":25,"_230":25,"_231":25,"_232":58,"_233":-5,"_234":-7,"_235":-7,"_236":137,"_237":25,"_238":25,"_239":-5,"_240":25,"_241":25,"_242":25,"_243":-5},"ENV",{"_195":196,"_197":198,"_199":86,"_200":201,"_202":203,"_204":205,"_206":207,"_208":209,"_210":211,"_212":213,"_214":215,"_216":217,"_218":219,"_220":221,"_222":223},"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 · v6