local-coder
github.com/gmarland/local-coder- Category
- AI Agents
- Rank
- No. 1786Tools index
Previous survey · No. 1695 ·
- Pricing
- Open Source
- Type
- AGENT
- Use case
- Agent Building
- GitHub
- 10 stars
- Date
About
local-coder builds a private, verified team of coding agents on top of local Ollama models and OpenCode. It detects the host machine's hardware, recommends models that fit, assigns them to seven specialist roles, and generates a permission-restricted workflow, then tests whether the resulting agents can actually use tools, delegate work, edit a repository, and verify the edit rather than trusting an agent's own report of success.
At a glance
local-coder configures a team of coding agents around Ollama and OpenCode, selecting models for the machine and probing whether they can do the work. It is worth trying if you want local inference but still need to verify edits and tool use.
Good fit
Developers with suitable local hardware who want a repeatable setup for several agent roles.
Consider something else when
Use a single local model if you only need occasional assistance and already have a working configuration.
Start here
Review the hardware recommendations and generated permissions. In a scratch repository, test one edit and check that the verifier can see the actual change and validation result.
Cost and requirements
MIT-licensed. The manifest requires Node.js 20 or later. The workflow uses Ollama and OpenCode, with model memory and storage needs depending on the selection.
Limits to know
A model that fits can still fail to call tools or complete edits. Review configuration changes and repeat capability checks when changing models or hardware.
Sources checked 2026-11-05. Source-based assessment.
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.