- I used my OpenAI API key.
- I used GPT-5.6-Terra.
- I tried to use GPT-5.6-Luna but it just complained about context window and that it cannot work with it.
So even though it is defenitely cheaper model, and it is astonishing value for money. But for me it just hit this issue.stream disconnected before completion: Request too large for gpt-5.6-luna in organization org-someid on tokens per min (TPM): Limit 200000, Requested 228484. The input or output tokens must be reduced in order to run successfully. Visit https://platform.openai.com/account/rate-limits to learn more.
- I tried to use GPT-5.6-Luna but it just complained about context window and that it cannot work with it.
- I set the reasoning to medium for the whole thing.
| extensions: | |
| health_check: | |
| endpoint: 0.0.0.0:13133 | |
| receivers: | |
| otlp: | |
| protocols: | |
| http: | |
| endpoint: 0.0.0.0:4318 | |
| grpc: |
I introduced a new API for synthesizing. But I wonder if the OTel logs will be traceable in synthesize.ts since it is a background job unlike speak.ts which is a forground job (tested, all the logs when a 3rd party service calls it with a trace ID attached to the request in Jaeger are still linked to the original reques. AKA its parent is set correctly).
Please make sure the requests coming from external services are traceable and we do NOT loss track of the request, this needs to be true even if we send multiple requests to the synthesize API.
Please also fix the issue of starting the app. I pasted the logs to this message.
So you must run the app, test first if the requests are traceable. And if they are do NOTHING, just explain to me why it is still even though they are running in the background. And if it is not please write a
plan.mdso I can see what needs to be done. >
...
The coding guidelines for this project are maintained in CONTRIBUTING.md. Please read and follow that file.
- Comments only when necessary.
- Be concise, short README, no emojis.
| { | |
| "name": "Business Development", | |
| "nodes": [ | |
| { | |
| "parameters": { | |
| "pollTimes": { | |
| "item": [ | |
| { | |
| "mode": "everyMinute" | |
| } |
| { | |
| "name": "Lead Hunter MCP", | |
| "nodes": [ | |
| { | |
| "parameters": { | |
| "path": "6bf0134b-d69a-420a-8d2c-8da8899a3ed9" | |
| }, | |
| "type": "@n8n/n8n-nodes-langchain.mcpTrigger", | |
| "typeVersion": 2, | |
| "position": [ |
| { | |
| "name": "My workflow", | |
| "nodes": [ | |
| { | |
| "parameters": { | |
| "options": {} | |
| }, | |
| "type": "@n8n/n8n-nodes-langchain.chatTrigger", | |
| "typeVersion": 1.4, | |
| "position": [ |
| { | |
| "name": "Stock Portfolio Agent", | |
| "nodes": [ | |
| { | |
| "parameters": { | |
| "rule": { | |
| "interval": [ | |
| { | |
| "field": "cronExpression", | |
| "expression": "0 12 * * 1-5" |
Ollama exposes an OpenAI-compatible API on port 11434, so most OpenAI-compatible tools can talk to it by simply changing the base URL. I am gonna try Qwen3-Coder 14B, light enough for a much nicer developer experience while still having enough resources left for Docker, VS Code, PostgreSQL, browsers, etc.
Ollama supports OpenAI-style /v1/chat/completions, you can test it:
curl http://localhost:11434/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
"model": "qwen3-coder:14b",
"messages": [