# AtomicBot-ai/atomic-agent

Local First Ai Agent. Optimized for Local Ai models. Long context window. Proper tools callings. Runs privately on your device.

Repository: https://github.com/AtomicBot-ai/atomic-agent
Canonical: https://ross.abutalabs.com/products/atomic-agent
Homepage: https://atomicagent.io
Language: TypeScript
License: MIT
License Family: permissive
Topics: agents, ai-agent, browser-automation, cli, gguf, llama-cpp, llm, local-first, nodejs, playwright, tui, typescript, hermes-agent, local-llm, openai, openclaw
Last push: 2026-08-26T20:54:00+00:00

## Health v2 (maintenance only)
Score: 77/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 87, longevity 9
- inputs: {"age_days": 134, "days_push": 7, "days_rel": 9, "gap_med": 0, "n_releases_24m": 66}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2446, forks 235 (observed 2026-08-28T04:06:51.616891+00:00)

## What it is
Atomic Agent is a local-first AI agent that runs entirely on your machine, using quantized models via llama.cpp to browse the web, edit files, run shell commands, remember context across sessions, and call tools over MCP. It is driven through a terminal UI, CLI, OpenAI-compatible HTTP API, or Telegram, with state kept in a local SQLite store.

## Use cases
- run an AI agent privately on my laptop without sending data to the cloud
- automate browser tasks with a local LLM
- have an agent read and edit files and run shell commands locally
- use small quantized GGUF models for long multi-step agent work
- embed an AI agent in my app via an OpenAI-compatible HTTP API
- control my computer from the terminal with an AI assistant
- avoid per-token API fees by running inference locally

## When to choose
- you need privacy-first, local-only agent execution with no cloud round-trips
- you want to run capable agents on consumer hardware with small quantized models
- you need reliable structured tool calls from local models (grammar-constrained decoding)
- you want a hackable, MIT-licensed agent you can extend with skills and MCP servers

## When to avoid
- you need frontier-model reasoning quality that only large cloud models provide
- you have no suitable local hardware (GPU/RAM) for running quantized models
- you need a stable, production-frozen API — the project is a developer preview with moving behavior
- you need Intel Mac support, which is not currently built

## Facets
- artifact type: cli-tool
- maturity: active
- function: agent-framework, llm-inference, cli, terminal-ui, http-server, browser-extension, mcp, chatbot
- domain: artificial-intelligence, large-language-models, developer-tools, privacy, self-hosted
- platform: windows, cli, cross-platform
- tags: local-first, llama-cpp, gguf, playwright, browser-automation, tauri-sidecar, quantized-models, openai-compatible-api, telegram, sqlite-memory, tool-calling, gbnf-grammar-constrained-decoding, ai-agents, command-line, automation, macos, linux, nodejs, desktop

## Member repositories
- AtomicBot-ai/atomic-agent (main) score 77

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:51.616891+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:30:31.861787+00:00, confidence not recorded.
  - readme: https://github.com/AtomicBot-ai/atomic-agent (fetched 2026-08-28T04:06:51.616891+00:00, sha d8d955639b91)
  - homepage: https://atomicagent.io (fetched 2026-08-29T10:11:56.729925+00:00, sha c7dd3142a8d6)
  - site_page: https://atomicagent.io/docs (fetched 2026-08-29T10:11:56.732881+00:00, sha 6d6fe80c0c06)
- Data as of 2026-08-30T08:39:29.467469+00:00.