The 2.5-flash-lite model was incorrectly matching the 2.5-flash
case, setting minimal budget to 128. Per Google's docs, flash-lite's
minimum thinking budget is 512, not 128. This caused the error:
The thinking budget 128 is invalid. Please choose a value between
512 and 24576.
Add a dedicated 2.5-flash-lite case before the 2.5-flash case so
the more specific match takes priority.
Fixes#2838
Co-authored-by: JasonOA888 <JasonOA888@users.noreply.github.com>
Google's promptTokenCount includes cachedContentTokenCount, so using it
directly as the input token count causes double-counting when calculateCost
multiplies input by the input rate AND cacheRead by the cacheRead rate.
The google-gemini-cli provider already handles this correctly (subtracting
cachedContentTokenCount from promptTokenCount), but google.ts and
google-vertex.ts were using the raw promptTokenCount.
This fix aligns both providers with the google-gemini-cli behavior.
When Google providers return tool calls without an args field (common for
no-argument tools), the arguments field was undefined. This breaks
subsequent API calls that require tool_use.input to be present.
Now defaults to {} when args is missing.
Related: clawdbot/clawdbot#1509
- Added headers field to base StreamOptions interface
- Updated all providers to merge options.headers with defaults
- Forward headers and onPayload through streamSimple/completeSimple
- Bedrock not supported (uses AWS SDK auth)
- Store thoughtSignature on text blocks during streaming (all 3 providers)
- Replay textSignature as thoughtSignature in convertMessages
- Remove redundant conditional since retainThoughtSignature handles undefined
Per Google docs, text part signatures are optional but recommended for
high-quality reasoning in multi-turn conversations.
- isThinkingPart now only checks thought === true, not thoughtSignature
- thoughtSignature is for context replay and can appear on any part type
- Store thoughtSignature on text blocks as textSignature for proper replay
- Remove id from functionCall/functionResponse (unsupported by Vertex/Cloud Code Assist)
Refs: https://ai.google.dev/gemini-api/docs/thought-signatures
Co-authored-by: Amp <amp@ampcode.com>
Google streaming may emit thoughtSignature without thought=true (including empty-text signature-only parts). Treat non-empty thoughtSignature as thinking to avoid leaking reasoning into normal text and retain signature across streaming deltas. Add unit test coverage.
- Add GoogleThinkingLevel type mirroring Google's ThinkingLevel enum
- Update GoogleGeminiCliOptions and GoogleOptions to use our type
- Cast to any when assigning to Google SDK's ThinkingConfig
- Add new API type 'google-cloud-code-assist' for Gemini CLI / Antigravity auth
- Extract shared Google utilities to google-shared.ts
- Implement streaming provider for Cloud Code Assist endpoint
- Add 7 models: gemini-3-pro-high/low, gemini-3-flash, claude-sonnet/opus, gpt-oss
Models use OAuth authentication and have sh cost (uses Google account quota).
OAuth flow will be implemented in coding-agent in a follow-up.
Previously, when using 'google-generative-ai' API with a custom baseUrl
in models.json, the baseUrl was ignored and requests always went to the
default Google endpoint.
Now the provider correctly passes model.baseUrl to the SDK's
httpOptions.baseUrl, enabling use of custom endpoints or API proxies.
Fixes#216
- Fix tool result format for Gemini 3 Flash Preview compatibility
- Use 'output' key for successful results (not 'result')
- Use 'error' key for error results (not 'isError')
- Per Google SDK documentation for FunctionResponse.response
- Improve type safety in google.ts provider
- Add ImageContent import and use proper type guards
- Replace 'as any' casts with proper typing
- Import and use Schema type for tool parameters
- Add proper typing for index deletion in error handler
- Add comprehensive test for Gemini 3 Flash tool calling
- Tests successful tool call and result handling
- Tests error tool result handling
- Verifies fix for issue #213Fixes#213
* use the correct Gemini 3 Flash Preview thinking levels
* fix a build error
* add changelog entry
* regenerate models
* make less assumptions about future models
- Added totalTokens field to Usage interface in pi-ai
- Anthropic: computed as input + output + cacheRead + cacheWrite
- OpenAI/Google: uses native total_tokens/totalTokenCount
- Fixed openai-completions to compute totalTokens when reasoning tokens present
- Updated calculateContextTokens() to use totalTokens field
- Added comprehensive test covering 13 providers
fixes#130
Fixes#39
- Added headers field to Model type (provider and model level)
- Model headers override provider headers when merged
- Supported in all APIs:
- Anthropic: defaultHeaders
- OpenAI (completions/responses): defaultHeaders
- Google: httpOptions.headers
- Enables bypassing Cloudflare bot detection for proxied endpoints
- Updated documentation with examples
Also fixed:
- Mistral/Chutes syntax error (iif -> if)
- process.env.ANTHROPIC_API_KEY bug (use delete instead of = undefined)
- Add NO_IMAGE to error finish reasons in Google provider
- Fix non-null assertion after optional chaining in Anthropic provider
- Migrate biome config to 2.3.5
- Ignore Tailwind CSS file from biome checks
- Bump all packages to version 0.6.0
Tool results now use content blocks and can include both text and images.
All providers (Anthropic, Google, OpenAI Completions, OpenAI Responses)
correctly pass images from tool results to LLMs.
- Update ToolResultMessage type to use content blocks
- Add placeholder text for image-only tool results in Google/Anthropic
- OpenAI providers send tool result + follow-up user message with images
- Fix Anthropic JSON parsing for empty tool arguments
- Add comprehensive tests for image-only and text+image tool results
- Update README with tool result content blocks API
Fixes issue where unpaired Unicode surrogates in tool results cause JSON serialization errors in API providers, particularly Anthropic.
- Add sanitizeSurrogates() utility function to remove unpaired surrogates
- Apply sanitization in all provider convertMessages() functions:
- User message text content (string and text blocks)
- Assistant message text and thinking blocks
- Tool result output
- System prompts
- Valid emoji (properly paired surrogates) are preserved
- Add comprehensive test suite covering all 8 providers
Previously only Google and Groq handled unpaired surrogates correctly.
Now all providers (Anthropic, OpenAI Completions/Responses, Google, xAI, Groq, Cerebras, zAI) sanitize text before API submission.
- Add 'aborted' as a distinct stop reason separate from 'error'
- Change AssistantMessage.error to errorMessage for clarity
- Update error event to include reason field ('error' | 'aborted')
- Map provider-specific safety/refusal reasons to 'error' stop reason
- Reorganize utility functions into utils/ directory
- Rename agent.ts to agent-loop.ts for better clarity
- Fix error handling in all providers to properly distinguish abort from error
- Switch from Zod to TypeBox for tool parameter schemas
- TypeBox schemas can be serialized/deserialized as JSON
- Use AJV for runtime validation instead of Zod's parse
- Add StringEnum helper for Google API compatibility (avoids anyOf/const patterns)
- Export Type and Static from main package for convenience
- Update all tests and documentation to reflect TypeBox usage
- Reduce default max tokens for Anthropic to 1/3 of model max
- Fix Google provider to properly handle empty tools array
- Ensure toolConfig is undefined when no tools are present
- Replace JSON Schema with Zod schemas for tool parameter definitions
- Add runtime validation for all tool calls at provider level
- Create shared validation module with detailed error formatting
- Update Agent API with comprehensive event system
- Add agent tests with calculator tool for multi-turn execution
- Add abort test to verify proper handling of aborted requests
- Update documentation with detailed event flow examples
- Rename generate.ts to stream.ts for clarity
- Google provider sometimes returns duplicate or missing tool call IDs
- Added counter to ensure unique IDs for each tool call
- Check for duplicates and generate new ID when needed
- Fixes issues with multiple tool calls having the same ID
- Updated generate-models.ts to fetch these providers directly from models.dev API
- OpenRouter now only used for xAI and other third-party providers
- Fixed test model IDs to match new model names from models.dev
- Removed unused import from google.ts
- Add getApi() method to all providers to identify the API type
- Add api field to AssistantMessage to track which API generated it
- Update transformMessages to check API compatibility instead of model
- Fixes issue where OpenAI Responses API failed when switching models
- Preserves thinking blocks and signatures when staying within same API
Mario Zechner