Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
Python agentaianthropicclaude-codecompressioncontext-engineeringcontext-windowcursorfastapilangchainllmmcpopenaiprompt-engineeringproxypythonragtoken-optimizationtokenstypescript
Commits on main
fix(learn): show prior patterns block to LLM to prevent dangling refs
6d2aba8 Garm committed 7 months ago
test(backpressure): harden livez-under-backpressure against CI jitter
924076a Garm committed 7 months ago
Merge pull request #239 from gglucass/pr/turn-id-tracking
2e351c6 Tejas Chopra committed 7 months ago
fix(subscription): raise default OAuth usage poll interval to 300s
a1d9832 chopratejas committed 7 months ago
Merge pull request #237 from gglucass/fix/traffic-learner-evidence-count
e4b8720 Tejas Chopra committed 7 months ago
Merge pull request #238 from omthorat2004/fix/mcp-cli-command
247060c Tejas Chopra committed 7 months ago
Merge pull request #227 from JerrettDavis/feat/canonical-pipeline
1fd8760 Tejas Chopra committed 7 months ago
docs(changelog): record turn_id feature; fix import order in new test
e8835af Garm committed 7 months ago
feat(proxy): emit turn_id linking agent-loop API calls from one user prompt
c4464c0 Garm committed 7 months ago
Merge pull request #233 from Kayzo/fix/dashboard-stats-snapshot-cache
4c9c29c Tejas Chopra committed 7 months ago
Merge pull request #230 from JerrettDavis/fix/copilot-oauth-runtime
e71ebd5 Tejas Chopra committed 7 months ago
Merge pull request #232 from gglucass/fix/learn-preserve-prior-recommendations
40f11d8 Tejas Chopra committed 7 months ago