Anvil
Active·★ 18·MIT·Updated 2026-10-12
Anvil is an open-source, self-hosted AI coding agent that provides a provider-agnostic LLM development pipeline. It allows developers to build, review, and ship code using various models across different stages, optimizing costs by routing tasks to the cheapest or most suitable provider for each step.
#AI Agent#LLM Orchestration#Code Generation#Multi-model#Self-hosted#Cost Management#DevOps#Observability
01
Features
01Provider-agnostic LLM pipeline orchestration with per-stage model routing and auto-failover.
02Automated multi-pass pull request review with evidence gates, incident binding, and GitHub integration.
03Long-term project memory with five types (working, episodic, semantic, procedural, profile) and code-fact drift detection.
04Advanced knowledge base with AST-aware chunking, hybrid retrieval (vector + BM25 + graph + rerank), and project graph.
05Live cost ledger per LLM call and OpenTelemetry observability for detailed tracking without estimates.
02
Compatibility
Node.js
Runtime
Verified via docs
TypeScript
Language
Verified via docs
Python
Language
Verified via docs
Go
Language
Verified via docs
Ollama
Local LLM
Verified via docs
OpenAI
LLM Provider
Verified via docs
03
Quick start
1
$ npm install -g @esankhan3/anvil-cli
2
$ anvil init
3
$ anvil dashboard
04
Use cases
↳Building multi-stage AI-powered software development pipelines with fine-grained cost optimization.
↳Automating pull request reviews and ensuring code quality across multiple repositories.
↳Conducting read-only code research and investigation, ensuring tasks remain in the free or cheapest tier.
↳Managing and leveraging long-term project knowledge and codebase conventions for AI agents.
↳Debugging and fixing bugs with a targeted, cost-effective retry loop using local or cheap LLM tiers.
05
Alternatives
refact★ 3.5k
AI Agent that handles engineering tasks end-to-end: integrates with developers’ tools, plans, executes, and iterates until it achieves a successful result.
mini-swe-agent★ 8.7k
The 100 line AI agent that solves GitHub issues or helps you in your command line. Radically simple, no huge configs, no giant monorepo—but scores >74% on SWE-bench verified!
docs-mcp★ 125
Add real-time chat, voice, video, and moderation to your app: Cursor's agent reads CometChat documentation, fetches reference pages, and pulls curated implementation bundles for React, React Native, Flutter, iOS, Android, the JavaScript SDK, the no-code widget, moderation, multi-tenant chat, and presence.
goal-to-game★ 104
Build beautiful games with high-quality 3D assets using Thrixel and Claude Code.
Comments
Log in to leave a comment
No comments yet. Be the first!