Uncaught Error: error sending request for url

[knightmarehs]

The agents can't generate conversations, the traceback is as follows:

8/12/2024, 5:33:59 PM [CONVEX A(aiTown/agentOperations:agentGenerateMessage)] Uncaught Error: error sending request for url (http://127.0.0.1:11434/api/embeddings): error trying to connect: tcp connect error: Connection refused (os error 61)

[ianmacartney]

check that you have something running on http://127.0.0.1:11434/api/embeddings - e.g. ollama - and that it's accessible from where you're running the backend

…

On Mon, Aug 12, 2024 at 6:44 AM knightmarehs ***@***.***> wrote: The agents can't generate conversations, the traceback is as follows: 8/12/2024, 5:33:59 PM [CONVEX A(aiTown/agentOperations:agentGenerateMessage)] Uncaught Error: error sending request for url (http://127.0.0.1:11434/api/embeddings): error trying to connect: tcp connect error: Connection refused (os error 61) — Reply to this email directly, view it on GitHub <#249>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACZQWYU5ZCM6MUVZLBZOI3ZRC34PAVCNFSM6AAAAABMMG4SEOVHI2DSMVQWIX3LMV43ASLTON2WKOZSGQ3DCMBVGQYDGOA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[uannyao]

Hi there, I'm facing a similar issue, the agents can't generate conversations, but in the traceback, it's saying the request was forbidden:

9/29/2024, 2:53:34 PM [CONVEX A(aiTown/agentOperations:agentGenerateMessage)] Uncaught Error: Request to http://localhost:11434/api/embeddings forbidden

[likeUMR]

I have the same problem too!!!

Uncaught Error: Request to http://127.0.0.1:11434/api/embeddings forbidden

When i change the embeddings to embed(base on a post about the ollama new api doc, so it is http://127.0.0.1:11434/api/embed) the problem still exist!

I have tested the api in terminal and it seems correctly working!
（curl -X POST http://127.0.0.1:11434/api/embed :11434/api/embed
-H "Content-Type: application/json"
-d '{"model":"mxbai-embed-large","input":"Stella is talking to Lucky"}'）

I don't know how to set the convex to fix that problem?

[quanchentg]

@likeUMR did you resolve this issue or figure something out? Thanks.

[quanchentg]

@ianmacartney is there any clue for this issue? Thanks.

[ianmacartney]

Sorry I haven't seen this issue before. I suggest dropping in logs, reproducing simple requests, verifying some assumptions (is ai-town running on your machine or in the cloud environment? is it running in a container? is ollama running somewhere accessible? You might also try 127.0.0.1 instead of `localhost` since some Windows users have seen odd behavior there. Best of luck with the debugging

…

On Sat, Oct 19, 2024 at 10:34 PM quanchentg ***@***.***> wrote: @ianmacartney <https://github.com/ianmacartney> is there any clue for this issue? Thanks. — Reply to this email directly, view it on GitHub <#249 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACZQW5A3HMBIF6YUMN4HIDZ4M6GJAVCNFSM6AAAAABMMG4SEOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIMRUGYYDQNZXGQ> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[leol15]

For me this happened when the frontend is pointing to Convex backend managed by Convex in the cloud (dashboard.convex.dev).

Check your .env.local file and make sure VITE_CONVEX_URL=http://127.0.0.1:3210 is there. If not you can re-generate it with npm run dev:backend after clearing the contents of .env.local.

Then you might need to restart the frontend app - npm run dev:frontend to pick up the change.

[quanchentg]

@leol15 Thanks for your nice suggestions. Let me give it a try!

[Tsailj]

@leol15 您好，想问您是使用本地的convex后端，不使用云端的convex就能成功跟Ollama通信是吗，我现在遇到的问题就是跟Ollama的连接被拒绝，我现在的convex使用的是云端的dashboard.convex.dev，.env.local也是设置的云端的没有用http://127.0.0.1:3210，您是使用http://127.0.0.1:3210就能避免通信被拒这个问题是吗

[Tsailj]

Uncaught Error: Request to http://127.0.0.1:11434/api/embeddings forbidden

When i change the embeddings to embed(base on a post about the ollama new api doc, so it is http://127.0.0.1:11434/api/embed) the problem still exist!

Hello, I’m facing the same issue as you. I also changed ‘embedding’ to ‘embed’, but I’m still getting an error（Uncaught Error: Request to http://localhost:11434/api/embed forbidden
）. Have you found a solution? Thank you.

[leol15]

@Tsailj Yes, local backend worked for me. This is the relevant command to start the backend https://github.com/a16z-infra/ai-town/blob/main/Justfile#L31

I'm unsure how the cloud backend was able to communicate with Ollama running locally, appreciate if others could explain!