{"data":{"slug":"atomicbot-ai-atomic-agent","name":"atomic-agent","tagline":"Local First Ai Agent for private device execution with optimized local AI models and long context window","github_url":"https://github.com/AtomicBot-ai/atomic-agent","owner":"AtomicBot-ai","repo":"atomic-agent","owner_avatar_url":"https://avatars.githubusercontent.com/u/259533419?v=4","primary_language":"TypeScript","stars":2523,"forks":246,"topics":["agents","ai-agent","browser-automation","cli","gbnf","gguf","hermes-agent","llama-cpp","llm","local-ai-agent","local-first","local-llm","mcp","nodejs","openai","openclaw","playwright","tui","typescript"],"archived":false,"github_pushed_at":"2026-09-17T16:56:38+00:00","maintenance_label":"Very active","stars_delta_30d":825,"url":"https://www.graphcanon.com/tools/atomicbot-ai-atomic-agent","markdown_url":"https://www.graphcanon.com/tools/atomicbot-ai-atomic-agent.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/atomicbot-ai-atomic-agent","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=atomicbot-ai-atomic-agent","description":"Atomic Agent is a local-first AI agent. Runs open-weight models on your own machine via llama.cpp.","homepage_url":"https://atomicagent.io","license":"MIT","open_issues":36,"watchers":17,"ai_summary":"AtomicBot-ai/atomic-agent is an AI agent built to optimize local model usage on personal devices, prioritizing privacy and performance with features like a long context window and proper tool calls.","readme_excerpt":"## Quick Install\n\nmacOS / Linux:\n\n```bash\ncurl -fsSL https://atomicagent.io/install | sh\n```\n\nWindows (PowerShell):\n\n```powershell\nirm https://atomicagent.io/install.ps1 | iex\n```\n\nThe installer downloads the release archive, verifies the checksum, and installs the CLI plus support assets (`grammars/`, native prebuilds, and bundled `ripgrep`). Atomic Agent updates itself in place; after an update the TUI prompts you to restart. Outside the TUI, run `atomic-agent update` (or `atag update`) to check for a newer release and re-run the installer in place — `atomic-agent update --check` probes without installing, and `--version <tag>` pins a specific release. Only the installed binary can self-update; a dev checkout updates via git.\n\n> [!NOTE]\n> Developer preview. APIs, commands, config, and behavior are still moving, so pin a release if you need a stable integration point. Current builds: macOS (Apple Silicon), Linux x64 / arm64, and Windows x64.\n\n---\n\n## Requirements & Configuration\n\n<details>\n<summary><b>Requirements</b> (Node, llama-server, browser, git) + Linux notes</summary>\n\n- Node.js for development; release bundles ship as Node SEA binaries.\n- A reachable `llama-server`, either managed by `atomic-agent models` or launched externally.\n- Managed mode picks the GPU backend automatically: Metal on Apple Silicon, CUDA on Windows when `nvidia-smi` reports a supported driver (including the reworked driver 610+ headers) with Vulkan as the fallback, Vulkan on Linux, CPU when no GPU is usable. Managed local models ship for Linux x64 only; on arm64 point the agent at an external `llama-server` instead.\n- Chrome, Microsoft Edge, or another configured Chromium-family executable. Browser binaries are not bundled.\n- `git` for git tools.\n- macOS workflows may need Accessibility, Screen Recording, Automation, or Reminders permissions.\n\n**Linux notes:**\n- **Desktop tools** (install via your package manager): `ripgrep` (file search; bundled binary used when present), `xclip`/`xsel` (X11) or `wl-clipboard` (Wayland) for clipboard, `libnotify-bin` for notifications, `wmctrl` for window control (X11/XWayland only), `gio` (glib2) or `trash-cli` for `fs.trash`.\n- **Browser:** Chromium-family sandboxing can fail under some Linux setups (containers, certain kernels). If Chrome refuses to launch, run it with `--no-sandbox`.\n- **GPU acceleration (managed mode):** the backend always starts and falls back to CPU when no GPU driver is available. For GPU offload install a Vulkan driver. Intel/AMD: `mesa-vulkan-drivers` (+ `vulkan-loader`/`libvulkan1`); NVIDIA: the stock proprietary driver bundles its Vulkan ICD. Device auto-selected at start; override with `atomic-agent models use-device <auto|cpu|Vulkan0>`, inspect with `atomic-agent models devices`, or press `G` in the TUI Models tab. Multi-GPU: set `localModels.managed.tensorSplit` in `config.json` (e.g. `[3, 1]` for a 75%/25% layer split) to launch llama-server with `--split-mode layer --tensor-split` across every visible GPU; combine with `use-device Vulkan0,Vulkan1` to restrict which devices join the split.\n\n</details>\n\n<details>\n<summary><b>How long a reply may run</b> (output ceilings)</summary>\n\nCloud models are **uncapped by default** — the service applies the model's own maximum, so one turn can write a whole file. To bound it (spend, or a service that requires the field) set `maxOutputTokens` on the provider entry:\n\n```json\n\"llm\": { \"providers\": [{ \"id\": \"openrouter\", \"kind\": \"openrouter\", \"maxOutputTokens\": 64000 }] }\n```\n\nReasoning models spend that same budget on thinking, so a low ceiling can be used up before any answer appears.\n\nLocal models use `localModels.completionMaxTokens` (llama.cpp's `n_predict`, default `8192`). Set it to `0` for no cap — generation then stops at a stop token or when the context window fills. That knob bounds time and runaway loops, not memory: what your machine commits is decided at daemon start by the model and `--ctx-size`, and does not grow with the length","github_created_at":"2026-04-21T21:34:51+00:00","created_at":"2026-07-15T10:59:01.279293+00:00","updated_at":"2026-09-20T05:01:30.035621+00:00","categories":[{"slug":"ai-agents","name":"AI Agents","url":"https://www.graphcanon.com/categories/ai-agents","markdown_url":"https://www.graphcanon.com/categories/ai-agents.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/ai-agents"}],"tags":[{"slug":"agents","name":"agents"},{"slug":"ai-agent","name":"ai-agent"},{"slug":"browser-automation","name":"browser-automation"},{"slug":"cli","name":"cli"},{"slug":"gguf","name":"gguf"},{"slug":"hermes-agent","name":"hermes-agent"},{"slug":"llama-cpp","name":"llama-cpp"},{"slug":"llm","name":"llm"}],"trust":{"provenance":{"is_fork":false,"github_id":1217416768,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-09-20T05:01:27.719Z","maintenance":{"label":"Very active","score":96,"methodology":"github_public_v1","releases_90d":30,"days_since_push":2,"last_release_at":"2026-09-17T17:09:50Z","stars_delta_30d":825,"open_issues_delta_30d":13},"security_summary":{"status":"no_manifest","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-09-20T04:00:45.757Z","medium_count":0,"scan_profile":"mcp_manifest","critical_count":0}},"capability_facts":{"mcp":{"source":"package.json:@modelcontextprotocol/*","observed_at":"2026-09-20T05:01:28.771Z","server_manifest":true},"scan":{"source":"repo_scan","observed_at":"2026-09-20T05:01:28.771Z"},"has_cli":{"value":true,"source":"package.json:bin|scripts","observed_at":"2026-09-20T05:01:28.771Z"},"languages":{"value":["typescript","javascript"],"source":"github.language+package.json","observed_at":"2026-09-20T05:01:28.771Z"},"license_spdx":{"value":"MIT","source":"github.license","observed_at":"2026-09-20T05:01:28.771Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["When you require an AI agent that prioritizes privacy by running exclusively on your device","If you need an AI assistant that works well with local AI models, offering a seamless performance without internet dependency","When the project involves tasks benefiting from long context windows to retain more information in conversations"],"when_not_to_use":["In scenarios where cloud-based services are mandatory for compliance or security reasons","If your use case demands real-time updates and data sharing capabilities across devices, atomic-agent's local-first approach may be a limitation"],"source":"enrich:decision_facts","observed_at":"2026-07-17T12:14:22.064Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"atomic-agent is a local-first AI agent designed to run optimized AI models privately on personal devices with features including long context windows and proper tool calls."}]}}