Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

What is KB_ENDPOINT? #21

Open
globalexport opened this issue Mar 5, 2024 · 8 comments
Open

What is KB_ENDPOINT? #21

globalexport opened this issue Mar 5, 2024 · 8 comments

Comments

@globalexport
Copy link

Hi!

What service url is expected in KB_ENDPOINT to scrape? I do not find documentation about it.

Thank you!

@sponzillo
Copy link
Contributor

Hi, the KB (RAG chatbots) is actually only available on the SAAS version. It will come to the open source version for the end of April 🙏🏻

Andrea

@0xM1gu3l
Copy link

0xM1gu3l commented May 3, 2024

Hey @globalexport, the KB_ENDPOINT is used to point to the tiledesk-llm project.
Just clone it to your machine and build it using Docker or pull the images at Docker Hub.
Configured it today and it works like a charm!

@globalexport
Copy link
Author

Thanks for the update! Now I understand it's a kind of API for the a pinecone vector db.

@globalexport
Copy link
Author

It would be neat to see the tiledesk-llm one day working together with a local vector db like chroma or similar. Sometimes there are use cases which do not allow to store data on an external host.

@vi1world
Copy link

Hey @globalexport, the KB_ENDPOINT is used to point to the tiledesk-llm project. Just clone it to your machine and build it using Docker or pull the images at Docker Hub. Configured it today and it works like a charm!

Can you please share more details about your configuration? Thank you

@0xM1gu3l
Copy link

Hey @globalexport, the KB_ENDPOINT is used to point to the tiledesk-llm project. Just clone it to your machine and build it using Docker or pull the images at Docker Hub. Configured it today and it works like a charm!

Can you please share more details about your configuration? Thank you

Hey @vi1world, basically in your docker-compose file you need to add these lines under the service:

 tilellm:
    container_name: tilellm
    image: tiledesk/tiledesk-backend-llm:latest
    build: ./tiledesk-llm/
    ports:
      - "8000:8000"
    environment:
      - PINECONE_API_KEY={YOUR_API_KEY}
      - PINECONE_INDEX={YOUR_INDEX_NAME}
      - PINECONE_TEXT_KEY={TEXT_KEY}

You'll need a account on Pinecone to use the LLM
After that just replace the variables with your keys and you're good to go.

Note: The last time that I used this was 4 months ago so I don't know if it had any changes, but give it a try!

@0xM1gu3l
Copy link

0xM1gu3l commented Sep 29, 2024

Another thing that I forgot is that you'll need to setup the keys KB_ENDPOINT, KB_ENDPOINT_QA and KB_ENDPOINT_TRAIN on some of the environments of the service.

On the server environment, set these variables:

- KB_ENDPOINT={KB_URL}
- KB_ENDPOINT_QA={KB_URL}
- KB_ENDPOINT_TRAIN={KB_URL}

They receive a url that probably is http://localhost:8000/api. If it's not working with localhost try to use the ip from the tiledesk-llm server. Good luck!

@vi1world
Copy link

vi1world commented Oct 4, 2024

Another thing that I forgot is that you'll need to setup the keys KB_ENDPOINT, KB_ENDPOINT_QA and KB_ENDPOINT_TRAIN on some of the environments of the service.

On the server environment, set these variables:

- KB_ENDPOINT={KB_URL}
- KB_ENDPOINT_QA={KB_URL}
- KB_ENDPOINT_TRAIN={KB_URL}

They receive a url that probably is http://localhost:8000/api. If it's not working with localhost try to use the ip from the tiledesk-llm server. Good luck!

Thank you very much.
I set it up and the LLM service is running but when adding a new knowledge base item, it doesn't get added to pinecone. I'm thinking it's maybe related to the queue not being processed. Any ideas?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants