Skip to content

Conversation

@fzowl
Copy link
Contributor

@fzowl fzowl commented Oct 14, 2025

Adding contextual model support
Counting the tokens and batching
Tests


Note

Upgrades the VoyageAI embedding stack for performance and model coverage.

  • Adds token-aware batching and count_tokens, with per-model token limits and unified embedding flow (regular, contextual via contextualized_embed, and multimodal via multimodal_embed)
  • Introduces helpers for model-type detection and batch building; sets large batch size; removes default model; makes api_key and model required in types
  • Updates provider wiring (client init hints) and removes default model from VoyageAIProvider
  • Adds extensive tests covering standard/contextual/multimodal models, custom dimensions, batching behavior, token limits, and edge cases
  • Refreshes docs (EN/KO/PT-BR) to recommend voyage-4, adds advanced config examples and model overview, and updates RAG Tool config options

Written by Cursor Bugbot for commit 15f49d4. This will update automatically on new commits. Configure here.

cursor[bot]

This comment was marked as outdated.

cursor[bot]

This comment was marked as outdated.

cursor[bot]

This comment was marked as outdated.

fzowl added 3 commits October 27, 2025 13:50
Supporting contextual models
Extending the tests
Supporting contextual models
Extending the tests
@fzowl fzowl force-pushed the context_model_and_token_counting branch from b628e2e to cc41421 Compare October 27, 2025 12:50
cursor[bot]

This comment was marked as outdated.

Copy link

@cursor cursor bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This PR is being reviewed by Cursor Bugbot

Details

Your team is on the Bugbot Free tier. On this plan, Bugbot will review limited PRs each billing cycle for each member of your team.

To receive Bugbot reviews on all of your PRs, visit the Cursor dashboard to activate Pro and start your 14-day free trial.

Bug: Default Model Removal Breaks Existing Configurations

Removing the default value for the model field breaks existing code that relies on the previous voyage-2 default. Users who don't explicitly specify a model in their VoyageAI embedder configuration will now encounter validation errors when the provider is instantiated, whereas their code worked before this change.

lib/crewai/src/crewai/rag/embeddings/providers/voyageai/voyageai_provider.py#L17-L21

)
model: str = Field(
description="Model to use for embeddings",
validation_alias="EMBEDDINGS_VOYAGEAI_MODEL",
)

Fix in Cursor Fix in Web


@fzowl
Copy link
Contributor Author

fzowl commented Nov 24, 2025

@lorenzejay Can you please take a look?

Copy link

@cursor cursor bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This PR is being reviewed by Cursor Bugbot

Details

Your team is on the Bugbot Free tier. On this plan, Bugbot will review limited PRs each billing cycle for each member of your team.

To receive Bugbot reviews on all of your PRs, visit the Cursor dashboard to activate Pro and start your 14-day free trial.

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

Successfully merging this pull request may close these issues.

1 participant