-
Notifications
You must be signed in to change notification settings - Fork 16
[0.2.0v][Feature] OpenAI provider support for embeddings and LLM #42
Copy link
Copy link
Labels
coreRelated to core engineRelated to core engineenhancementNew feature or requestNew feature or request
Milestone
Description
Overview
Add OpenAI as an alternative AI provider alongside Gemini, so users who don't have a Gemini API key can use Simili-Bot with OpenAI instead.
Scope
- Auto-select provider based on available API keys (
GEMINI_API_KEYpreferred,OPENAI_API_KEYas fallback) - Support
OPENAI_API_KEYacross all workflow examples and docs - Update embedder and LLM client initialisation to be provider-agnostic
- Update
.env.sample,simili.yamlexamples, and setup guides to document both providers - Clarify supported embedding models and their dimensions per provider
Acceptance Criteria
- At least one of
GEMINI_API_KEYorOPENAI_API_KEYmust be set (validated at startup) - Gemini is preferred when both keys are present
- Default OpenAI models documented (
text-embedding-3-smallfor embeddings) - All existing tests pass with the new provider abstraction
- No breaking changes for existing Gemini-only users
Notes
Extracted from PR #40. Implementation reference: nick1udwig/simili-bot@index-and-query-prs.
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
coreRelated to core engineRelated to core engineenhancementNew feature or requestNew feature or request
Type
Projects
Status
Done