Skip to content

Release 0.1.2

Compare
Choose a tag to compare
@karolisg karolisg released this 21 Feb 14:43
· 6 commits to main since this release
7572288

Changelog

All notable changes to this project will be documented in this file. See standard-version for commit guidelines.

0.1.2 (2025-02-21)

Features

  • Add api_invoke spans (8398924)
  • Add clickhouse dependency (4e6ae44)
  • Add cost control and limit checker (17eab2c)
  • Add database span writter (94c048a)
  • Add extra to request (a1ff5fb)
  • Add missing gemini parameters (c22b37c)
  • Add model name and provider name to embeddings API (e1d365f)
  • Add rate limiting (459ba9d)
  • Add server crate with sample configuration (a2e7c90)
  • Build for ubuntu and docker images (#3) (1e29aad)
  • display models (8a1efdf)
  • Enable otel when clickhouse config provided (5cdadb1)
  • implement mcp support (90220c2)
  • Implement tui (#4) (7589219)
  • Improve UI (#15) (b83d183)
  • Integrate routed execution with fallbacks (#20) (3d75331)
  • Print provider and model name in logs (c8832e1)
  • Refactor targets usage for percentage router (6d04e2d)
  • Support .env variables for config (546d2a6)
  • Support langdb key (#21) (767e05e)
  • Support search in memory mcp tool (#29) (5d71a78)
  • Use in memory storage (bf35718)
  • Use time windows for metrics (#28) (c6ed8e4)
  • Use user in openai requests (68415b0)

Bug Fixes

  • Add router span (1860f51)
  • Create secure context for script router (3cc7b8a)
  • Fix connection to mcp servers (e2208f8)
  • Fix gemini call when message is empty (4a00a25)
  • Fix langdb config load (#26) (8f02d58)
  • Fix map tool names to labels in openai (436d09e)
  • Fix model name in models_call span (62f5a38)
  • Fix provider name (#18) (7fdc24a)
  • Fix provider name in tracing (e779ec7)
  • Fix response format usage (#22) (dbaf61d)
  • Fix routing direction for tps and requests metrics (b96ee3e)
  • Fix serialization of user properties (e8830c8)
  • Fix tags in tracing (0b1ae3e)
  • Fix tonic shutdown on ctrl+c (3c42dba)
  • Fix tracing for openai and deepseek (bbbae94)
  • Improve error handling in loading config (12a5cb2)
  • Return authorization error on invalid key (000c376)
  • Return formated error on bedrock validation (543585d)
  • Store inference model name in model call span (bbedf30)

0.1.1 (2025-02-21)

Features

  • Add api_invoke spans (8398924)
  • Add clickhouse dependency (4e6ae44)
  • Add cost control and limit checker (17eab2c)
  • Add database span writter (94c048a)
  • Add extra to request (a1ff5fb)
  • Add missing gemini parameters (c22b37c)
  • Add model name and provider name to embeddings API (e1d365f)
  • Add rate limiting (459ba9d)
  • Add server crate with sample configuration (a2e7c90)
  • Build for ubuntu and docker images (#3) (1e29aad)
  • display models (8a1efdf)
  • Enable otel when clickhouse config provided (5cdadb1)
  • implement mcp support (90220c2)
  • Implement tui (#4) (7589219)
  • Improve UI (#15) (b83d183)
  • Integrate routed execution with fallbacks (#20) (3d75331)
  • Print provider and model name in logs (c8832e1)
  • Refactor targets usage for percentage router (6d04e2d)
  • Support .env variables for config (546d2a6)
  • Support langdb key (#21) (767e05e)
  • Support search in memory mcp tool (#29) (5d71a78)
  • Use in memory storage (bf35718)
  • Use time windows for metrics (#28) (c6ed8e4)
  • Use user in openai requests (68415b0)

Bug Fixes

  • Add router span (1860f51)
  • Create secure context for script router (3cc7b8a)
  • Fix connection to mcp servers (e2208f8)
  • Fix langdb config load (#26) (8f02d58)
  • Fix map tool names to labels in openai (436d09e)
  • Fix model name in models_call span (62f5a38)
  • Fix provider name (#18) (7fdc24a)
  • Fix provider name in tracing (e779ec7)
  • Fix response format usage (#22) (dbaf61d)
  • Fix routing direction for tps and requests metrics (b96ee3e)
  • Fix serialization of user properties (e8830c8)
  • Fix tags in tracing (0b1ae3e)
  • Fix tonic shutdown on ctrl+c (3c42dba)
  • Fix tracing for openai and deepseek (bbbae94)
  • Improve error handling in loading config (12a5cb2)
  • Return authorization error on invalid key (000c376)
  • Store inference model name in model call span (bbedf30)