We are excited to announce the official release of OMNI v0.5.8! This release finalizes Phase 1 of our Context Analytics roadmap by introducing a robust, memory-efficient streaming architecture.

Streaming Distillation Pipeline

To handle long-running and piped commands without memory exhaustion, we’ve introduced a memory-efficient, line-by-line processing pipeline (src/hooks/pipe.rs). This ensures that even massive logs stream through OMNI seamlessly.

Expanded Semantic Distillers

We have added new declarative TOML signal profiles for 9 critical developer tools. OMNI now provides deep semantic understanding for: aws, az, bun, deno, docker, gcloud, npm, vite, and webpack.

Brand & Documentation Modernization

The English README.md and all 6 i18n localized versions have been completely overhauled! They now feature professional SVG visual assets (hero.svg, architecture.svg), and brand new “Under the Hood” and “Real-World Use Cases” sections.

Advanced Context Analytics

For deeper visibility into token reduction and signal density, we’ve implemented context composition metrics in omni stats (src/analytics/context_composition.rs).

Improved Storage & Hook Routing

SQLite Storage Enhancements: Updated src/store/sqlite.rs to seamlessly support high-throughput, chunked streaming outputs from the new distillation pipeline.
MCP & Tooling Synchronization: Refined pre_tool and post_tool hook routing to perfectly align with the new semantic models.