8 Commits

Author	SHA1	Message	Date
Debanjum	181332dcb8	Improve Claude context caching to improve response cost, intelligence ... Old thought messages are dropped by default by the Anthropic API. This change ensures old thoughts are kept. This should improve cache utilization to reduce costs. And keeping old thoughts may also improve model intelligence.	2025-12-07 20:23:58 -08:00
Debanjum	da493be417	Support image generation with Gemini Nano Banana	2025-11-29 15:57:35 -08:00
Debanjum	57d6ebb1b8	Support Google Gemini 3 ... - Use thinking level for gemini 3 models instead of thinking budget. - Bump google gemini library - Add default context, pricing	2025-11-18 14:13:24 -08:00
Debanjum	3496189618	Support using MCP tools in research mode ... - Server admin can add MCP servers via the admin panel - Enabled MCP server tools are exposed to the research agent for use - Use MCP library to standardize interactions with mcp servers - Support SSE or Stdio as transport to interact with mcp servers - Reuse session established to MCP servers across research iterations	2025-11-16 10:50:30 -08:00
Debanjum	69cceda9ab	Bump server dependencies	2025-11-11 16:08:38 -08:00
Debanjum	e854c1a5a8	Bump django, langchain python server dependencies	2025-09-16 14:03:58 +05:30
Debanjum	0a0b97446c	Avoid `click' v8.2.2 server dependency as it breaks pypi validation ... Refer pallets/click issue 3024 for details	2025-08-01 23:36:13 -07:00
Debanjum	006b958071	Use UV to install server for speed, package locks in dev setup, workflows ... It's much faster than pip, includes dependency locks via uv.lock and comes with standard convenience utilities (e.g pipx, venv replacement)	2025-08-01 00:01:34 -07:00