RLM — Recursive Language Model

\n External Context\n \n

\n \n

0 chars · ~0 tokens

\n\n

\n Query\n \n

\n\n

Samples

\n \n \n \n \n

\n\n

\n Depth\n \n \n \n \n \n

\n\n

\n \n \n

\n\n

RLM

\n Load context, enter a query, hit RUN.
\n The model explores context through tools, never seeing it directly.
\n Look for the ✦ Final Answer at the bottom.\n

\n\n\n\n`;\n\n// ─── HTTP Handler ───\n\nexport default async function (req: Request): Promise {\n const url = new URL(req.url);\n\n if (req.method === \"GET\" && (url.pathname === \"/\" || url.pathname === \"\")) {\n return new Response(HTML, {\n headers: { \"Content-Type\": \"text/html; charset=utf-8\" },\n });\n }\n\n if (req.method === \"POST\" && url.pathname === \"/run\") {\n const { query, context, maxDepth } = await req.json();\n\n if (!query || !context) {\n return new Response(\n JSON.stringify({ error: \"query and context required\" }),\n {\n status: 400,\n headers: { \"Content-Type\": \"application/json\" },\n },\n );\n }\n\n const depth = Math.min(Math.max(maxDepth || 2, 1), 5);\n\n // Stream SSE response\n const { readable, writable } = new TransformStream();\n const writer = writable.getWriter();\n\n // Run in background, close stream when done\n (async () => {\n try {\n await runRLM(query, context, 0, depth, writer);\n await sseEvent(writer, \"done\", {});\n } catch (err: any) {\n await sseEvent(writer, \"error\", { depth: 0, error: err.message });\n } finally {\n await writer.close();\n }\n })();\n\n return new Response(readable, {\n headers: {\n \"Content-Type\": \"text/event-stream\",\n \"Cache-Control\": \"no-cache\",\n \"Connection\": \"keep-alive\",\n \"Access-Control-Allow-Origin\": \"*\",\n },\n });\n }\n\n return new Response(\"Not found\", { status: 404 });\n}","deleted","label",[100],"permalink","https://www.val.town/x/curtcox/ptc-rlm/code/main.ts","extension","actionData","errors"]