Search tool now supports 4 backends with unified output (bocha,
qianfan, zhipu, linkai) and a routing layer:
- strategy 'auto' (default): pick first configured in canonical order
bocha > qianfan > zhipu > linkai
- strategy 'fixed': pin a specific provider
- agent may pass `provider` to override per-call (only exposed when
≥2 providers configured + auto strategy)
- Unified flush + context injection into a single async LLM call
(flush_from_messages accepts context_summary_callback)
- Fixed response parsing bug: handle generator returns and Claude-format
dicts from bot.call_with_tools, which previously caused all LLM
summaries to silently fail (falling back to rule-based extraction)
- Removed standalone context summary prompts and methods; reuse the
existing [DAILY]/[MEMORY] summarization pipeline
- Updated docs (zh/en/ja) to reflect the new injection behavior
Browser tool enhancements:
- Navigate action now auto-includes snapshot result, saving one LLM round-trip
- Wait for networkidle + 800ms after navigation for SPA/JS-rendered pages
- Prompt guides agent to screenshot key results and ask user for login/CAPTCHA help
- Fixed playwright version pinned to 1.52.0; mirror fallback to official CDN on failure
Web console file/image support:
- SSE real-time push for images and files via on_event (file_to_send)
- Added /api/file endpoint to serve local files for web preview
- Frontend renders images in media-content container (survives delta/done overwrites)
- File attachment cards with download links; RFC 5987 encoding for non-ASCII filenames
Tool workspace fix:
- Inject workspace_dir as cwd into send and browser tools (previously only file tools)
- Screenshots now save to ~/cow/tmp/ instead of project directory
- Rename ZHIPU_AI bot type from glm-4 to zhipu to avoid confusion with model names
- Add bot_type persistence in web config to fix provider dropdown resetting on refresh
- Change OpenAI provider key to chatGPT to match bot_factory routing
- Add DEEPSEEK constant and route it to ChatGPTBot (OpenAI-compatible API)
- Keep backward compatibility for legacy bot_type glm-4 in bot_factory
- Fix re.sub bad escape error on Windows paths by using lambda replacement
- Remove unused pydantic import in minimax_bot.py
Made-with: Cursor
- Use LLM to summarize discarded context into concise daily memory entries
- Batch trim to half when exceeding max_turns/max_tokens, reducing flush frequency
- Run summarization asynchronously in background thread, no blocking on replies
- Add daily scheduled flush (23:55) as fallback for low-activity days
- Sync trimmed messages back to agent to keep context state consistent
- Fix dark mode FOUC: apply theme in <head> before first paint, defer
transition-colors to post-init to avoid animated flash on load
- Fix Safari IME Enter bug: defer compositionend reset via setTimeout(0)
- Fix history scroll: use requestAnimationFrame before scrollChatToBottom
- Limit restore turns to min(6, max_turns//3) on restart
- Fix load_messages cutoff to start at turn boundary, preventing orphaned
tool_use/tool_result pairs from being sent to the LLM
- Merge all assistant messages within one user turn into a single bubble;
render tool_calls in history using same CSS as live SSE view
- Handle empty choices list in stream chunks
- Simplify skill-creator installation flow
- Refine skill selection prompt for better matching
- Add parameter alias and env variable hints for tools
- Skip linkai-agent when unconfigured
- Create skills/ dir in workspace on init
