other / agent

Apache Maka (Incubating)

An agent workspace with desktop, terminal UI and CLI clients, built around an append-only log of model messages, tool calls and permission decisions.

Repository owner apache

Not rated yet
0 comments

Like 0 Save

Published 11 Oct 2026Listing updated 12 Oct 2026

At a glance

Purpose: An agent workspace with desktop, terminal UI and CLI clients, built around an append-only log of model messages, tool calls and permission decisions.
Runs in: Local agent with shell
Operating systems: macOS, Linux, Windows
Needs: Runtime: Node.js, npm, Git and ripgrep; Needs internet: Not stated; Needs accounts or API keys beyond your agent: Not stated
Files and data: Reads files: Yes; Writes files: Yes; Sends data to other services: Not stated
Tool type: Agent
Delivery: From GitHub
Version: Latest on the main branch
Access: Code on GitHub
Compatibility: According to the README, Maka lists macOS as a platform, with Windows and Linux in preview. Desktop uses Electron. You can connect a cloud API, a local model or a compatible gateway. The README mentions Model Context Protocol client integration, a computer-use package and a public npm package for the CLI. It names 0.2.0 (incubating) as the latest Apache release.
External costs: According to the README, Maka does not bundle a shared model account. You bring your own model: a cloud API, a local model or a compatible gateway. The README says a successful resume of an interrupted turn calls the model and uses tokens. The repository is Apache-2.0-licensed and states no pricing. The model provider you connect may charge for usage.
Permissions: The README shows npm dependency installation for source builds and says development commands start Electron with hot reload. The README says workspace data, settings, connection metadata, plaintext credentials and artifacts are stored under Electron userData. The README says file-writing and shell tools pass through a sandbox boundary and permission decisions are recorded in the event log. The README lists MCP integration, computer-use backends and a ripgrep-based Grep tool. The README says graph implementation operators create isolated Git worktrees and require a clean source worktree. The README describes model API, local-model and account connections supplied by the user.
Usage terms: Apache License 2.0, as stated in the repository. See its LICENSE file for the exact terms. The README points to NOTICE for attribution information and says third-party components remain subject to their respective licences and notices.
Interface: desktop

Code on GitHub

apache/maka

Apache-2.0

TypeScript5,706 stars545 forksLast pushed 13 Oct 2026 UTC

Apache Maka (Incubating) is a high-performance agent workspace that keeps a complete record of everything it did.

Topics: agent-runtime ai ai-agent apache cli desktop electron event-sourcing

Public GitHub data fetched 13 Oct 2026 UTC · 0 days old

View on GitHub ↗

What it does

Apache Maka (Incubating) is an agent workspace. The README says Desktop, TUI, CLI and Eval use one Runtime Host, and that users supply a cloud API, local model or compatible gateway.

The README describes model messages, tool calls, permission decisions and termination as append-only RuntimeEvents, with the UI, next prompt and crash recovery derived from the log. The README identifies runtime.sqlite as the live record and describes explicit interrupted-turn resume.

The README's repository layout lists an Electron desktop app, terminal clients, MCP integration, computer-use backends, evaluation code and shared UI. The README says graph mode uses --graph in the CLI or /graph on in the TUI, and its implementation operators work in isolated Git worktrees.

Who it is for

Developers who want a local agent workspace with desktop, terminal and command-line clients and an append-only event log.

Example tasks

Add a model connection in Settings → Models and choose a default model.
Start the terminal UI from a source checkout with npm run cli:dev.
Run one task from the command line with npm run cli:dev -- run "Summarize this repository and identify its most important risk".
Use /resume in the CLI or TUI to resume an interrupted turn.
Run a graph task from the CLI with the run command's graph option, from a clean Git worktree.

Getting started

According to the README, the official Apache release is a signed source archive, and packages distributed elsewhere are convenience artifacts. It links a downloads page and a CLI installation guide for the public npm package. To run from a source checkout, clone the repository, then run npm ci and npm run dev to start Desktop with hot reload. For the terminal UI, run npm run build and then npm run cli:dev. On first launch, the README says to open Settings → Models, add a connection, use its test step and choose a default model. The README has the exact commands.

Limitations

The README requires Node.js 22.19+ on the 22 line or Node.js 24+, npm, Git and ripgrep for source builds; the README says Node.js 23 is unsupported and CI uses Node.js 24. The README says Direct Peer and Peer Mesh development additionally require Rust stable 1.98+ and the platform linker, naming Xcode Command Line Tools on macOS and MSVC Build Tools on Windows. The README lists macOS, Windows preview and Linux preview. The README says users must add a model connection and that credentials are stored in local plaintext. The README says graph implementation needs a clean Git worktree. The README says older JSONL transcripts and Electron safeStorage credentials are not imported, so upgraded workspaces may show empty threads and require credentials to be entered again. The README labels the project incubating and not fully endorsed by the ASF, distinguishes development builds from approved Apache releases, and warns that formats, commands and experimental capabilities may change. The repository can change at any time; check its README before you install.

Data practices

The README says sessions, settings and run records are stored locally under Electron userData. The README lists runtime.sqlite, settings.json, connection-catalog.json, credential-vault.json and an artifacts folder. The README describes credential-vault.json as local plaintext for API keys and similar secrets. The README says users supply a cloud API, local model or compatible gateway. The README does not detail what model context is transmitted to those connections. The README links SECURITY.md and a workspace privacy document for further details. The README does not mention telemetry.