This website requires JavaScript.
Explore
Help
Sign In
redtea
/
lpra
Watch
1
Star
0
Fork
0
Code
Issues
Pull Requests
Actions
Packages
Projects
Releases
Wiki
Activity
Files
03f963340e3b099ed154e56cdcde2c6dc549e826
lpra
/
internal
T
History
redtea
03f963340e
feat: «Думать» uses a model's own thinking when it has it (OpenAI-compatible reasoning, Anthropic extended thinking) instead of asking it to think in words too; learned per model, the word-thinking prompt kept for models without; without «Думать» such a model is asked not to think
2026-10-17 15:36:35 +03:00
..
agent
docs: http-lp forwarding goes over the agent's WebSocket, by long polls only without one - said so in the panel, the code and the README
2026-10-16 21:56:24 +03:00
proto
docs: http-lp forwarding goes over the agent's WebSocket, by long polls only without one - said so in the panel, the code and the README
2026-10-16 21:56:24 +03:00
server
feat: «Думать» uses a model's own thinking when it has it (OpenAI-compatible reasoning, Anthropic extended thinking) instead of asking it to think in words too; learned per model, the word-thinking prompt kept for models without; without «Думать» such a model is asked not to think
2026-10-17 15:36:35 +03:00
tunnel
fix: small fixes
2026-10-13 19:00:50 +03:00