Releases: openintelligence-labs/tokenmiser
Release list
TokenMiser v0.6.1
Full Changelog: v0.6.0...v0.6.1
TokenMiser v0.6.0
What's Changed
- v0.6.0: streaming cache, request coalescing, budget limits, security hardening by @thinkflow19 in #1
New Contributors
- @thinkflow19 made their first contribution in #1
Full Changelog: v0.5.1...v0.6.0
TokenMiser v0.5.1
Patch release: the dashboard now works fully offline, in line with the ecosystem's local-first / no-external-fetch principle.
Fixed
- Dashboard fully offline — htmx 2.0.4 is vendored into the binary and served from the local
/assets/htmx.jsroute instead of a CDN<script>tag. The vendored file is integrity-verified: byte-identical across unpkg and jsdelivr, and its SHA-384 matches htmx's officially published SRI hash for 2.0.4. Regression tests assert the dashboard HTML and live fragment contain no external http(s) URLs.
Full details in CHANGELOG.md.
Install (macOS arm64)
curl -sL https://github.com/openintelligence-labs/tokenmiser/releases/download/v0.5.1/tokenmiser-v0.5.1-aarch64-apple-darwin.tar.gz | tar xz
./tokenmiser-v0.5.1-aarch64-apple-darwin/tokenmiserOther platforms: cargo build --release from source.
SHA-256 (tokenmiser-v0.5.1-aarch64-apple-darwin.tar.gz): cfe92395468bc1a179e1aa2e0596e25ef4cea0372304c2c2ce44fb9cd5bbe57d
Full Changelog: v0.5.0...v0.5.1
TokenMiser v0.5.0
First feature release: TokenMiser is now a full Rust LLM cost-routing proxy — a 9-crate workspace built on Pingora.
Highlights
- Drop-in OpenAI-compatible proxy —
/v1/chat/completions, streaming (SSE) and non-streaming, with per-response audit headers (x-tokenmiser-cache,x-tokenmiser-tier,x-tokenmiser-difficulty,x-tokenmiser-routed-to) - Tiered router — Tier 0 heuristics → Tier 1 semantic classifier → Tier 2 speculative cascade, plus a Rhai policy DSL with
tokenmiser policy testreplay - Dual-layer semantic cache — L1 exact-match LRU + L2 semantic (bge-small embeddings, precision-tuned 0.87 cosine threshold)
- Provider adapters — OpenAI, Anthropic, Ollama behind one trait; local-first: auto-detects a running Ollama and routes easy traffic there with zero API keys and zero config
- Cost ledger + live dashboard — real-time USD spent/saved from a canonical
pricing.json; htmx dashboard at/and JSON/stats - Quality judge — shadow-mode A/B with LLM-as-judge win-rate auto-gate
- MCP budget gateway — per-agent/per-tool spend caps via
/v1/mcp/*
No telemetry, ever. Full details in CHANGELOG.md.
Install (macOS arm64)
curl -sL https://github.com/openintelligence-labs/tokenmiser/releases/download/v0.5.0/tokenmiser-v0.5.0-aarch64-apple-darwin.tar.gz | tar xz
./tokenmiser-v0.5.0-aarch64-apple-darwin/tokenmiserOther platforms: cargo build --release from source.
SHA-256 (tokenmiser-v0.5.0-aarch64-apple-darwin.tar.gz): 888c0e8e861ea7f142d0f72a25a0b73147ccc11a04c782388871183792a21a3e