Is this another LiteLLM?

Same job, narrower promise: a self-hosted Rust fast path — one binary plus Postgres, no Redis — for OpenAI and Anthropic-compatible traffic with Model Groups, budgets, and usage records.

How do I route Jev-style decisions through it?

Point System One traffic at /v1/systemone or /v1/decisions. They serve Jev-compatible backends such as Ollaya and Laya or hosted Jev-style endpoints.

Are the benchmark numbers real?

They are honest about scope: local Rust mocks measuring router overhead — single-digit milliseconds at 1M tokens — not model inference, with artifacts committed per scenario.

What does it store about my traffic?

Usage metadata in Postgres: never prompt text, uploads, tool payloads, or answers. Team budgets, key expiry, and rate limits sit on top.

How do I install and start it?

A one-line install script clones or updates the repository at $HOME/brighto-router, creates .env when missing, pulls the official Docker image, starts PostgreSQL, runs migrations, seeds defaults and starts the router. It needs Docker with Compose, Bash, Python 3.9+, Git and curl, and the Portal opens on port 18080.

What capabilities does the router expose?

OpenAI-compatible chat, completions and Responses; Anthropic Messages; embeddings, rerank and ASR pass-through; System One and decisions at /v1/systemone and /v1/decisions for TypeSafe or Jev-compatible models; Model Groups for round-robin or weighted balancing; and teams, client keys, expiry, rate and concurrency limits, and a usage dashboard.

How do Model Groups and budgets behave?

Model Groups put same-type routes behind one API model name using round-robin or weighted round-robin, and a failed endpoint is skipped after repeated pre-response failures and retried after BACKEND_CIRCUIT_OPEN_SECONDS, default 30. Token budgets block requests when reached, but a configured money amount is only stored and shown in the Portal, not enforced against actual spend.

CommunityJev Integrations & Adapters

BrighTO_Router

Self-hosted Rust LLM gateway and load balancer with a System One decision route: one endpoint for OpenAI, Anthropic, local, and Jev-compatible backends, with Model Groups, team budgets, and Postgres usage records.

See highlights

Open github.com

Stars: 21
Language: Rust
License: APACHE-2.0
Published by: Community
Author: thusinh1969
Last verified: Oct 2026

thusinh1969/BrighTO_Router

User repository on GitHub

View on GitHub

BrighTO LLM Router: free open-source, ultra-fast Rust LLM gateway with a single-file Python SDK, OpenAI/Anthropic APIs, SystemOne decisions for Ollaya/Laya, Quyết and JEV/DJEV, model load balancing, fallback routing, team API keys and token budgets.

Forks: 5
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.Homepage

About this project

Million-token AI traffic, simple Rust fast path, one Docker install.

Quoted from the project's README. Read the full README.

Tagscommunity github

Works forJev for Routing

Used inJev for Software & DevTools

Languages: Rust

Highlights

/v1/systemone and /v1/decisions routes serve Jev-compatible decision models such as Ollaya/Laya or hosted Jev-style endpoints.
Model Groups spread one API name across compatible routes with round-robin or weighted balancing; router overhead benchmarks in single-digit milliseconds.
One Docker image plus Postgres with no Redis; records usage metadata but never prompt text. One-line install.

Watch out

Benchmarks measure router overhead on local mocks, not model inference; Linux and Docker only. Copy reviewed against the repo README on 2026-10-04, not from running it.

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

Holon

159

Rust agent workbench for ongoing work that keeps branch decisions in their own crate and backend notes in the adapter. Its first decision backend is Cloudflare's Clef, borrowing Clef's System One-shaped state and questions request instead of reinventing the interface.

Jev Integrations & Adapters#community#agent-runtime#rust

Community

holon-run/holonRust

Astra

Self-hosted agent runtime that adds an optional Jev judgment layer to memory filtering, lesson dismissal, request classification, and skill selection, with inspectable context and execution.

Jev Integrations & Adapters#community#jevbest#catalog-import

Community

matrixorigin/AstraRust

Turn

Programming language and VM where agents run as isolated actors with managed context, durable effects, and confidence-aware control flow, including a Jev decision effect.

Jev Integrations & Adapters#community#jevbest#catalog-import

Community

ekizito96/TurnRust

jev-ai-use-cases

LangChain notebook with runnable Jev examples for ticket triage, model routing, reply guardrails, tool selection and finance inbox checks.

Jev Integrations & Adapters#community#github

Community

atliq/jev-ai-use-casesJupyter Notebook

opencode-auto-jev

OpenCode plugin adding an Auto virtual model that asks Jev to pick one of your connected models per turn, failing closed to a fallback when unsure.

Jev Integrations & Adapters#community#github

Community

jorgefspereira/opencode-auto-jevTypeScript

ai-chess-lab

Browser chess lab where LLMs, System One models, and humans play single games, Swiss tournaments, and leagues with Stockfish analysis and cost tracking.

Jev Integrations & Adapters#community#github

Community

dperezcabrera/ai-chess-labPython

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai