
The Nano Collective is fiscally hosted by Open Source Collective. Every incoming and outgoing payment appears on the public OSC ledger as it happens, not as a yearly report. The numbers are not a promise, they are a URL anyone can watch. https://nanocollective.org



The collective is volunteer-led. Bounties sit alongside that work, not above it. For any paid piece: scope, deliverables, and amount are written into the issue before code lands. No retrospective deals. https://nanocollective.org


get-md's `--compare` runs the same URL through both conversion paths (fast Turndown, local LLM) and writes the two outputs side by side in one file. Pick the one you want, no rerun. https://github.com/Nano-Collective/get-md


`migrate()` and `migrateAsync()` throw three distinct error types: `ValidationError` (output vs schema), `MigrationError` (your `up()` threw), `VersionError` (bad config). Branch on cause, not a generic catch. https://github.com/Nano-Collective/json-up



Before Nanocoder touches a file: Shift+Tab into plan mode. The AI investigates the codebase and writes up a structured plan (files, steps, risks) without editing anything. You review it, then switch back and execute.

https://github.com/Nano-Collective/nanocoder



Nanotune fine-tunes small models on Apple Silicon without you writing a config file. Each step is an interactive prompt. MLX and llama.cpp are wired up underneath, so you can go from data to GGUF without learning the stack.

https://github.com/Nano-Collective/nanotune
