new-api
Open-source AI gateway. Self-hosted LLM gateway with OpenAI-, Claude- and Gemini-compatible APIs, model routing, access control, usage tracking and cost accounting.
- 52k
- Go
- AGPL-3.0
- 1 d ago
Open-source AI gateway. Self-hosted LLM gateway with OpenAI-, Claude- and Gemini-compatible APIs, model routing, access control, usage tracking and cost accounting.
Local Anthropic-compatible AI gateway with cheap-first routing, quality gates, safe model handoffs, and an embedded cost dashboard.
Automatic LLM router — 82% cost savings, 79.4% accuracy, 93.4% pass rate. Drop-in OpenAI proxy.
Source-available LLM router and AI cost optimizer. Routes simple prompts to cheap or local models and complex ones to premium models, automatically. Drop-in OpenAI-compatible proxy for Claude Code, Codex, Cursor and OpenClaw. Self-hosted, no middleman.
A curated, source-backed list of projects built with Jev, TypeSafe AI's System One model for typed decisions.
Make your AI agents work as a team. You keep working in 1 session while it orchestrates models via your other subscriptions, local models and more. Run a Software Factory: long plans, proven steps, safe merges. Carries your setup, tracks every cost.
Cheap Money-Saving Self-hosted AI agent workspace with tool calling, MCP, multi-model routing, sandboxed execution, multi-agent workflows, and LLM-authored 3D character animation rendered with Three.js.
Open-source AI CLI and local MCP server connecting 25 clients—Claude Code, Cursor, Codex, ChatGPT, Hermes, and OpenClaw—to 2,000+ models/APIs, with OAuth and rollback.
Engineering deterministic, production-grade systems around non-deterministic LLMs — FSM, durable execution, retries, DAGs, agent runtimes, model routing, edge inference, RAG, memory, multi-agent orchestration, security, and observability. 14 runnable proof-of-concept phases.
ChatGPT 模型路由检测器:查看服务器报告的真实响应模型,自动识别路由降级,GPT Pro 计划降智查询。|ChatGPT Model Route Inspector: view the actual response model reported by the server, automatically detect routing downgrades, and check whether GPT Pro requests are being downgraded.
Route complex Codex tasks to model-specific background workers with bounded concurrency and lead-agent verification.
Turn Fable or Opus into your agent orchestrator: plan coding work, assign capable Claude, Codex, or Grok agents, and personally verify the result.
Private AI assistant, AI agent, and unified model proxy for Claude Code, Codex CLI, Gemini CLI & OpenClaw. Skills, MCP, tools, channels, tasks,model routing, accounts, keys, logs, dashboard.
Provider-pluggable orchestration runtime for multi-model AI inference. ( Sakana Fugu style )
Self-hosted, Docker-first OpenAI-compatible LLM gateway with provider routing, fallbacks, API keys, and usage controls.
Drop-in LLM router and AI gateway for OpenAI Chat Completions, Responses and Anthropic Messages APIs: 125 routing strategies (incl. feedback-aware Thompson sampling) across GPT-6, Claude Opus 5.5 / Fable 5.2, Gemini 4 and Kimi K3.1, a request feedback API, budgets, fallbacks, tenant rate limits, idempotency and audit logs.
A curated list of strategies, tools, papers, and resources for reducing LLM token costs and improving efficiency in production.
Nexus — execution infrastructure for AI agents. Host agents in the cloud or behind NAT; let users attach and authorize their own computers and Android phones. Deploy once, work across devices. Python SDK · MCP · Shared run context · Browser, terminal and files · Built-in chat UI · Model routing.
Auto-delegation, status truth, and cost-aware model routing for OpenClaw agents
Claude Code plugin that picks the model and effort for each turn (micro, low, medium, or high tier), advised by Jev, Clef, OpenAI, or a local Ollama classifier.
Make every Claude token count. rightsize picks the right model and effort for each prompt, while keeping cache and context costs low. It learns how you work and gets better over time.
Claude Code and Codex CLI skill: an LLM router that delegates bulk, already-decided coding tasks to another model you already pay for (GPT via Codex, Gemini via agy, a Claude Opus subagent), on another subscription first. Benchmark-checked routes, cost per task measured on a bench, cross-model review. The host verifies every diff.
A prompt-aware LLM router that predicts which models can complete each request, then selects the cheapest capable one: 53.2% lower cost and +1.9 pts completion on our tested dataset.
Local-model (Ollama) integration for DeepSeek Harness: discover, pull, remove, and inspect local models, route requests to them by task type or keyword with automatic fallback to the cloud, and get a one-shot status overview via /ollama.
Per-step LLM routing benchmark with 970 static labels, live SWE-bench evaluation, an open data pipeline, and a public leaderboard.
Quality-per-token GPT-5.6 Sol, Terra, and Luna model router for Codex, Work, CLI, and subagents.
ORBIT: A Unified Toolbox for LLM Routing. Unified LLM/model routing, cost-aware inference, multimodal benchmarks, and reproducible evaluation. Accepted by Frontiers of Computer Science (2026).
Open-source, model-portable coding agent built for cost-efficient, verifiable software work—with model routing, progressive context, compact tool output, explicit task state, and estimated savings.
Route every coding task to the best AI agent on your machine — Claude Code, Codex, Cursor, Gemini CLI, Aider, OpenCode. Installs as a Claude Code plugin, Gemini extension, or Agent Skill.
Governed neural weight editing and coding model adaptation. Native Rust patches, exact rollback, LoRA, Claude Code routing, and reproducible benchmarks.
Open-source LLM routing evaluation platform benchmarking cost, latency, quality, and safety across routing strategies, with a focus on Turkish banking.
Self-hosted model routing gateway and agent control plane for OpenAI, Anthropic, Gemini and A2A agents behind one OpenAI-compatible API, protocol-first. MCP server security, governed tools, prompt enhancement, virtual-key budgets, cache, observability, audit chain and agent governance. Zero-key local first run; public preview.
[FROZEN 2026-10-31] Automatic strong/cheap model-tier routing for DeepSeek Harness: intent-gated tier landing on the agent/request waterfall, plan-mode handoff (strong plans, cheap implements), deterministic high-risk guards on tools/pre-execute, failure escalation with TTL fallback, a /tier command, and tier status tools
A blind panel of models + a judge, for decisions where being wrong is expensive. Verifies before it synthesizes; counts consensus by model family, not by head.
Text classification model, used with transformers.
Text classification model, used with transformers.
Runs in the browser on Hugging Face