Skip to content

feat(documents): support Gemini native document input - #110

Draft
binsarjr wants to merge 4 commits into
anvie:devfrom
binsarjr:feat/native-document-analysis-gemini
Draft

feat(documents): support Gemini native document input#110
binsarjr wants to merge 4 commits into
anvie:devfrom
binsarjr:feat/native-document-analysis-gemini

Conversation

@binsarjr

@binsarjr binsarjr commented Aug 13, 2026

Copy link
Copy Markdown
Contributor

What this adds

analyze_document can send supported files to Gemini through its native generateContent endpoint. The document is passed inline with its MIME type, while the user's question and the existing untrusted-document system prompt stay intact.

This also adds the Google Gemini provider preset for both new and existing installations. In this adapter, PDF and RTF are enabled; unsupported binary Office formats are skipped so fallback routing can choose another compatible model.

Review note

Depends on #108. Until that PR is merged, GitHub will also show its parent commits here; the Gemini-specific change is the top commit.

Verification

  • Gemini and document-analysis tests: 22 passed
  • GitHub Python 3.11 suite: all 3 Gemini tests passed; 2,176 passed and 95 skipped overall, with the same 2 existing dev failures
  • Go tests passed
  • Python compilation and git diff --check passed

@binsarjr
binsarjr force-pushed the feat/native-document-analysis-gemini branch from 7f76ba1 to 00dcbee Compare August 13, 2026 10:45
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