@columenlabs/lumen-ai 0.1.147 → 0.1.148

package/dist/assistant/engine.d.ts

@@ -240,6 +240,12 @@ export declare class Assistant {
     /** The loop on Anthropic's Messages API, with the advisor and prompt caching. */
     private loopClaude;
     /** Swaps the history for a short summary. False with nothing to compact. */
+    /**
+     * One plain answer from OpenAI, metered on `s`: no tools, no chat. Streamed and read from the raw
+     * events like the chat loop, because a ChatGPT plan sign-in answers only streamed calls, takes no
+     * output cap and may send its words only as deltas.
+     */
+    private openaiOnce;
     private compact;
     /**
      * One prompt answered on a task's tier as the person: no chat, no tools. It is for background

package/dist/assistant/engine.js

@@ -1697,6 +1697,59 @@ export class Assistant {
         return answer;
     }
     /** Swaps the history for a short summary. False with nothing to compact. */
+    /**
+     * One plain answer from OpenAI, metered on `s`: no tools, no chat. Streamed and read from the raw
+     * events like the chat loop, because a ChatGPT plan sign-in answers only streamed calls, takes no
+     * output cap and may send its words only as deltas.
+     */
+    async openaiOnce(clients, s, instructions, input, max, signal) {
+        const client = clients.openai;
+        const plan = !!clients.openaiPlan;
+        let text = '';
+        let done;
+        try {
+            const stream = await client.responses.create({
+                model: s.model.id,
+                instructions,
+                input,
+                store: false,
+                stream: true,
+                ...(plan ? {} : { max_output_tokens: max }),
+            }, signal ? { signal } : {});
+            for await (const e of stream) {
+                if (e.type === 'response.output_text.delta' && typeof e.delta === 'string')
+                    text += e.delta;
+                else if (e.type === 'response.completed' ||
+                    e.type === 'response.incomplete' ||
+                    e.type === 'response.failed')
+                    done = e.response;
+                else if (e.type === 'error') {
+                    const er = (e.error ?? e);
+                    throw new AssistantError(openaiStreamError(er.code || er.type || '', plan), 'upstream');
+                }
+            }
+        }
+        catch (err) {
+            // The SDK throws a stream's `error` event as an APIError with no HTTP status.
+            if (err instanceof OpenAI.APIError && err.status === undefined) {
+                const er = (err.error ?? {});
+                throw new AssistantError(openaiStreamError(er.code || er.type || err.code || '', plan), 'upstream');
+            }
+            throw err;
+        }
+        if (done?.status === 'failed')
+            throw new AssistantError(openaiStreamError(done.error?.code ?? '', plan), 'upstream');
+        const u = done?.usage;
+        if (u) {
+            const read = u.input_tokens_details?.cached_tokens ?? 0;
+            s.meter.add({
+                input_tokens: Math.max(0, u.input_tokens - read),
+                output_tokens: u.output_tokens,
+                cache_read_input_tokens: read,
+            }, plan);
+        }
+        return text || done?.output_text || '';
+    }
     async compact(clients, s, doc, emit) {
         if (doc.history.length < 2)
             return false;
@@ -1706,21 +1759,7 @@ export class Assistant {
             .join('\n\n');
         let summary = '';
         if (s.model.provider === 'openai' && clients.openai) {
-            const res = await clients.openai.responses.create({
-                model: s.model.id,
-                instructions: system,
-                input: chat,
-                store: false,
-                max_output_tokens: 1500,
-            });
-            const u = res.usage;
-            if (u)
-                s.meter.add({
-                    input_tokens: u.input_tokens - (u.input_tokens_details?.cached_tokens ?? 0),
-                    output_tokens: u.output_tokens,
-                    cache_read_input_tokens: u.input_tokens_details?.cached_tokens ?? 0,
-                });
-            summary = res.output_text ?? '';
+            summary = await this.openaiOnce(clients, s, system, chat, 1500);
         }
         else if (clients.anthropic) {
             const msg = await clients.anthropic.messages.create({
@@ -1764,21 +1803,7 @@ export class Assistant {
         let text = '';
         try {
             if (s.model.provider === 'openai' && clients.openai) {
-                const res = await clients.openai.responses.create({
-                    model: s.model.id,
-                    instructions: o.system,
-                    input: o.prompt,
-                    store: false,
-                    max_output_tokens: max,
-                }, o.signal ? { signal: o.signal } : undefined);
-                const u = res.usage;
-                if (u)
-                    s.meter.add({
-                        input_tokens: u.input_tokens - (u.input_tokens_details?.cached_tokens ?? 0),
-                        output_tokens: u.output_tokens,
-                        cache_read_input_tokens: u.input_tokens_details?.cached_tokens ?? 0,
-                    }, !!clients.openaiPlan);
-                text = res.output_text ?? '';
+                text = await this.openaiOnce(clients, s, o.system, o.prompt, max, o.signal);
             }
             else if (clients.anthropic) {
                 const msg = await clients.anthropic.messages.create({

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@columenlabs/lumen-ai",
-  "version": "0.1.147",
+  "version": "0.1.148",
   "description": "Lumen's assistant and agent parts: a chat assistant's server (Claude tiers, tools over MCP, the web, hand off, spend caps), serving an MCP server from a Worker, read-only tool sets and metered tool calls.",
   "license": "MIT",
   "type": "module",
@@ -41,8 +41,8 @@
   },
   "dependencies": {
     "@anthropic-ai/sdk": "0.133.0",
-    "@columenlabs/lumen-ops": "0.1.147",
-    "@columenlabs/lumen-ui-types": "0.1.147",
+    "@columenlabs/lumen-ops": "0.1.148",
+    "@columenlabs/lumen-ui-types": "0.1.148",
     "openai": "7.32.0",
     "zod": "^4.3.6"
   },