[H-337] Fix Captain provider runtime and knowledge flow (#61)
* fix(H-337): configure Captain provider runtime * fix(captain): make knowledge rebuild atomic * fix(captain): scope retrieval provider failures --------- Co-authored-by: Rogee <rogee@ipao.vip>
This commit is contained in:
@@ -47,6 +47,20 @@ Default login after seeding:
|
||||
|
||||
Change these values in `.env` before running the seed command if needed.
|
||||
|
||||
## Optional Captain provider
|
||||
|
||||
Set `GOCHAT_COPILOT_PROVIDER_CONFIG` and `GOCHAT_COPILOT_CHAT_API_KEY` in the
|
||||
ignored `.env` to enable Captain chat and embeddings. The JSON uses the same
|
||||
`chat`, `embedding`, `generation`, and `request` contract as the SuperAdmin
|
||||
Copilot provider settings API. Set `GOCHAT_COPILOT_EMBEDDING_API_KEY` only when
|
||||
`embedding.mode` is `separate`; runtime secrets are not written to the database.
|
||||
|
||||
Example settings value (the key remains a separate `.env` variable):
|
||||
|
||||
```json
|
||||
{"chat":{"provider":"openai_compatible","base_url":"https://provider.example.com/v1","model":"gpt-5.6-luna"},"embedding":{"mode":"reuse_chat_credentials","provider":"openai_compatible","base_url":"https://provider.example.com/v1","model":"text-embedding-3-small","dimensions":1536},"generation":{"temperature":0.7,"max_tokens":2048},"request":{"timeout_seconds":60,"max_retries":2}}
|
||||
```
|
||||
|
||||
## Useful Commands
|
||||
|
||||
```bash
|
||||
|
||||
Reference in New Issue
Block a user