How to use Maki: what you can do with it

3 use cases

• Install Maki and cut the tokens your coding agent spends

  Install the Maki coding agent and use its index and code execution tools to read less and do more per call.

  • REPO tontinton/maki
  • Software and DevOps
  • Beginner
  View prompt

  Install tontinton/maki, configure it with <provider>, and run it on <repo> to <task>. Report how many lines it read and which tools it used.

  Share Email me this

• Run a coding agent on your own Ollama or llama.cpp server

  Point Maki at a local model so code never leaves your machine and there is no API bill.

  • REPO tontinton/maki
  • Software and DevOps
  • Intermediate
  View prompt

  Set up tontinton/maki to use my local <Ollama or llama.cpp> server at <host>, pick the best coding model I have, and run a small refactor in <repo> to test it.

  Share Email me this

• Replace a headless Claude Code script with Maki

  Run Maki without the UI and keep the same stream-json output your scripts already parse.

  • REPO tontinton/maki
  • Claude Code
  • Software and DevOps
  • Intermediate
  View prompt

  Change my <script> that runs Claude Code in print mode to call tontinton/maki with --print --output-format stream-json instead, run it on <test task>, and diff the outputs.

  Share Email me this

What people are building with Maki

• Benchmarkthe web

  Ran Maki, Pi and OpenCode on one local 30B model and found Maki gave the strongest repo review at 43K context

  Same prompt, model and llama.cpp server on a 128 GB Strix Halo box: Maki took 511 seconds and 53 tool calls but peaked at 43.1K context thanks to indexed search and bounded reads, against 90.6K for OpenCode, and produced the strongest review, with Pi a close second.

  Maki on GitHubTry it: Run a coding agent on your own Ollama or llama.cpp server

  By Piotr Karbowski (linkedin.com) on the web·Aug 10, 2026

  View original post on the web

• Hands-onGitHub

  Used Maki with a local Qwen 27B in Ollama to rebase a GitLab patch and hit indentation errors Pi did not

  A user gave Maki, with rtk and qwen3 27b on Ollama, the task of rebasing a Rails patch from GitLab 18.x to 19.x. Repeated runs produced wrong indentation while Pi with the same model got it right; later runs without rtk were mixed but the problem still appeared.

  Maki on GitHub

  By @yuyuyureka on GitHub·Aug 21, 2026

  View original post on GitHub

Real builds shared on X, Hacker News, Reddit, DEV and GitHub, summed up by Deduped from each post and its links; each card links to the original post. See what people are building with other repos

Quick start with Maki

1. Check you have what it needs: Linux, macOS or Windows; An API key or sign-in for a supported provider, or a local Ollama or llama.cpp server; Optional: rtk installed for smaller bash output.
2. Open tontinton/maki on GitHub and read the README and license.
3. Download the installer with curl -fsSL https://maki.sh/install.sh -o install.sh and read it with cat install.sh
4. Run it with chmod +x install.sh && sh install.sh, or use brew install maki
5. Set your provider key, for example OPENAI_API_KEY or GEMINI_API_KEY
6. Start Maki in your project and give it a task; watch the index tool read skeletons before full files

These steps come from our top use case, Install Maki and cut the tokens your coding agent spends, which also has a copy-paste agent prompt.

Using the Deduped MCP, get the setup steps for tontinton/maki

claude mcp add --transport http deduped https://dedupedai.com/mcp

Best AI repos for Software and DevOps

About this repo

• Who it is forDevelopers who pay per token or run local models and want a fast coding agent that does not waste context.
• Why it is hot1.3k stars, talked about in 3 AI sources we track
• LanguageRust
• Sourcegithub.com/tontinton/maki · Back to Hot repos

Requirements

• Linux, macOS or Windows
• An API key or sign-in for a supported provider, or a local Ollama or llama.cpp server
• Optional: rtk installed for smaller bash output

Related repos

• Trended earlier REPO ggml-org/llama.cpp llama.cpp llama.cpp is a dependency-free C/C++ engine for running LLMs and vision language models locally or in the cloud, with quantization, many GPU backends, a CLI and an OpenAI-compatible server.
• Trended earlier REPO egoist/waku Waku Waku is a fast, native desktop app written in Rust that gives you one place to run and manage local coding agent CLIs, projects, and sessions.
• Trended earlier REPO tony3liu/eido Eido Eido is a macOS AI coding workspace, built on Zed and pi, for understanding code, working with an agent, and reviewing its changes.