Nanocoder
Nanocoder is a free terminal coding agent built by a volunteer community rather than a company — no paid tiers, no telemetry, and you choose whether it runs on your own local model or a cloud provider.
🔗 Visit Nanocoder
Description
Most AI coding agents are built by companies with a business model to protect — free tiers exist to funnel you toward a paid plan, and your prompts often go through the vendor's servers. Nanocoder takes a different path: it's built by the Nano Collective, a volunteer community rather than a company, with no paid tiers gating useful features and no telemetry quietly shipping your prompts anywhere.
It runs agentic coding sessions using the model of your choice — a fully local model via Ollama, or any OpenAI-compatible API (OpenRouter, Anthropic, Google, and others) — so you decide where your code and data go. It offers normal, auto-accept, yolo and plan modes, inline and fullscreen rendering, CLI support for non-interactive workflows, an MCP server integration, a skills system with subagents and tools, and session checkpointing. It's installable via npm, Homebrew or Nix Flakes and is entirely free and open-source.
💬 Our review
The short version: if you specifically want a terminal coding agent that's genuinely community-run rather than a company's funnel toward a paid tier, and you're comfortable choosing and configuring your own model backend, Nanocoder is a solid, honest option.
Against GitHub Copilot or Claude Code (both company-built, polished, but closed-source with vendor telemetry and pricing) Nanocoder's edge is transparency and control — no hidden data flows, no paywalled features. The trade-off is that a volunteer-run project won't have the same polish, support responsiveness, or guaranteed longevity as a well-funded commercial product, and you're responsible for supplying and configuring your own model access. Worth it for privacy-conscious developers who want full control over their coding agent stack; less appealing if you just want something that works out of the box with zero setup.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Entièrement gratuit et open-source, sans palier payant
Pros
Projet communautaire, sans télémétrie ni palier payant caché
Compatible modèles locaux (Ollama) ou API cloud au choix
Modes normal/auto-accept/yolo/plan, intégration MCP
Installation simple (npm, Homebrew, Nix)
Cons
Pas de support commercial garanti — projet bénévole
Nécessite de configurer soi-même l'accès au modèle
Moins poli/mature qu'un produit commercial financé