agent · AI Agents Frameworks

gptme

gptme is a free, open-source, terminal-based AI agent that runs anywhere a terminal runs, from a laptop to SSH sessions, tmux, headless servers, and CI pipelines. It ships with built-in tools for shell/Python execution, file editing, web browsing, and vision, supports multiple LLM providers (Anthropic, OpenAI, Google, xAI, DeepSeek, OpenRouter, or fully local via llama.cpp) with a swappable -m/--model flag, and has built-in MCP support plus multi-agent/concurrent-agent capability. It can run interactively with per-action confirmation prompts or in a fully non-interactive/autonomous mode (-n flag), and is deployable via pipx/uv, a self-hosted server+web UI, or a desktop app, with community support via Discord.

gptme.orgupdated Aug 2, 2026view as JSON

27 fields evidenced · 28 with no public information. Every value below links to the document it came from and the date we checked it.

Trust gap

How this is scored

Runtime governance

2/3

Are scope, spend, and authority enforced while the agent runs?

Tools are scoped by an explicit --tools allow-list and actions require confirmation by default, with a described guardrail stack; limits are user-configurable rather than centrally enforced.

Audit trail

0/3

Is there a tamper-evident record of what it actually did?

No public evidence found on the page checked: no audit trail or retained record is described.

How this gap gets closed →

Compliance

0/3

Are compliance obligations attached per engagement?

No public evidence found on the page checked: no certification or regulatory alignment is published.

How this gap gets closed →

Outcome settlement

0/3

Does payment depend on a verified result?

Free and open-source with bring-your-own provider keys; no outcome-linked settlement.

How this gap gets closed →

Insurance & indemnity

0/3

Can the deployment be insured, and is the customer indemnified?

No public evidence found on the page checked: no cover, carrier or indemnity is described.

How this gap gets closed →

Assurance

Insurance available: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Insurance carriers: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Coverage limits: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Indemnification: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Liability cap: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Audit trail: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Tamper-evident log: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Explainability: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Runtime governance: Describes a layered guardrail stack (selectors, lessons, checks) governing agent runtime behavior.
“This stack is simple and composable: selectors improve work choice,”
gptme.org · checked Aug 2, 2026
Permission scopes: Which tools/permissions are enabled can be scoped via the --tools flag (comma-separated allow-list).
“-t, --tools TEXT Tools to allow as comma-separated list. Available: append, browser, chats, choice, computer, gh, ipython, morph, patch, rag, read, save, screenshot, shell, subagent, tmux, vision.”
gptme.org · checked Aug 2, 2026
Compliance certifications: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Regulatory alignment: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Outcome-based pricing: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Settlement mechanism: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Dispute process: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
SLA terms: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Evaluation coverage: Has an evaluation suite for testing capabilities of different models.
“Evaluation suite for testing capabilities of different models.”
gptme.org · checked Aug 2, 2026

Agency

Autonomy level: Supports fully non-interactive/autonomous mode (-n flag) as well as interactive/confirm-required mode.
“# Fully non-interactive/autonomous mode (no user interaction possible, safe for scripts/CI) gptme -n 'run the test suite and fix any failing tests'”
gptme.org · checked Aug 2, 2026
Human oversight: Tool actions require confirmation prompts by default (-y flag skips confirmation).
“-y, --no-confirm Skip all confirmation prompts.”
gptme.org · checked Aug 2, 2026
Goal complexity: General-purpose agent for coding and other knowledge-work goals, not limited to narrow single-step tasks.
“A great coding agent, but general-purpose enough to assist in all kinds of knowledge-work.”
gptme.org · checked Aug 2, 2026
Action space: Built-in tools include shell, python, patch (file edits), browser, vision, tmux, morph, form, and mcp.
“gptme ships with comprehensive tools: shell : Execute commands in terminal python : Run Python code patch : Make incremental file edits browser : Browse web via Playwright vision : Process images/screenshots tmux : Manage tmux sessions morph : Fast file transformations form : Interactive forms mcp : MCP tool discovery and use”
gptme.org · checked Aug 2, 2026
Operating environment: Runs anywhere a terminal runs: laptop, SSH sessions, tmux, headless servers, CI pipelines.
“A personal AI agent that runs anywhere a terminal runs — your laptop,”
gptme.org · checked Aug 2, 2026
Initiative: Reference autonomous agent (Bob) proactively opens PRs and reviews code without step-by-step direction.
“Bob opens PRs, reviews code,”
gptme.org · checked Aug 2, 2026

Safety

Safety evaluations: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Red teaming: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Safety policy: Describes guardrails (input, pre-action, output) that persistent agents need around the full action loop.
“🛡 Guardrails Persistent agents need guardrails around the full loop, not just tool”
gptme.org · checked Aug 2, 2026
Usage restrictions: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Model or system card: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Incident reporting: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Third-party evaluations: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Data handling: Emphasizes local-first data handling — can run fully locally via llama.cpp so data, models, and execution stay on the user's terminal.
“or fully local via llama.cpp — your data, your models, your terminal.”
gptme.org · checked Aug 2, 2026

Practicality

Pricing model: No gptme subscription fee; users bring their own LLM API keys and pay providers directly, with a /tokens command to track usage and costs.
“/tokens Show token usage and costs”
gptme.org · checked Aug 2, 2026
Price point: The core CLI is free and open-source; users bring their own LLM provider API keys (cost is provider usage, not a gptme price).
“Free and open-source.”
gptme.org · checked Aug 2, 2026
Availability: Available as installable CLI, plus a desktop app for Linux, macOS, and Windows with auto-updates.
“Desktop app for Linux (AppImage), macOS, and Windows with auto-updates — download”
gptme.org · checked Aug 2, 2026
Deployment options: Deployable via pipx, uv tool install, with optional extras, or from git source with all extras.
“# With pipx (recommended, requires Python 3.10+) pipx install gptme # With uv uv tool install gptme # With optional extras pipx install 'gptme[browser]' # Playwright for web browsing pipx install 'gptme[all]' # Everything # Latest from git with all extras uv tool install 'git+https://github.com/gptme/gptme.git[all]'”
gptme.org · checked Aug 2, 2026
Integrations: Integrates as a drop-in coding agent in editors like Zed and JetBrains IDEs via ACP.
“This makes gptme available as a drop-in coding agent in Zed and JetBrains IDEs.”
gptme.org · checked Aug 2, 2026
Supported regions: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Support model: Community support via Discord for questions and discussion; no paid support tier described.
“Discord — ask questions, share what you've built, discuss features”
gptme.org · checked Aug 2, 2026

Foundation models

Base models: Supports Claude (Anthropic), GPT (OpenAI), Gemini (Google), Grok (xAI), DeepSeek, OpenRouter (100+ models), and local llama.cpp models.
“gptme supports multiple AI providers: Provider Setup Features Anthropic ANTHROPIC_API_KEY Claude models OpenAI OPENAI_API_KEY GPT models Google GEMINI_API_KEY Gemini models xAI XAI_API_KEY Grok models DeepSeek DEEPSEEK_API_KEY DeepSeek models OpenRouter OPENROUTER_API_KEY Multi-provider gateway llama.cpp Local server OpenAI-compatible local models”
gptme.org · checked Aug 2, 2026
Model provider: Provider-agnostic: 100+ models via OpenRouter/local, plus direct Anthropic, OpenAI, Google providers, per comparison table.
“Provider support ✅ 100+ models (OpenRouter/local) Anthropic/Bedrock/Vertex Anthropic/OpenAI/Google OpenAI (built-in)”
gptme.org · checked Aug 2, 2026
Model swappable: Model is swappable via -m/--model CLI flag (e.g. switching between OpenAI and Anthropic models).
“-m, --model TEXT Model to use, e.g. openai/gpt-5, anthropic/claude- sonnet-4-20250514. If only provider given then a default is used.”
gptme.org · checked Aug 2, 2026
Open weights: Supports fully local operation with local/open-weight models via llama.cpp (no API key required).
“Local models via llama.cpp need no key — see providers docs .”
gptme.org · checked Aug 2, 2026
Fine-tuning: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Context window: no public information
gptme.org · checked Aug 2, 2026 · source did not state this

Ecosystem

Protocols supported: MCP (Model Context Protocol) is built-in supported, per the feature comparison table.
“MCP Support ✅ Built-in ✅ Built-in ❌ ❌ ”
gptme.org · checked Aug 2, 2026
Tool use: Equipped with a rich set of built-in tools (shell, python, browser, vision, files, etc.).
“gptme equips the AI with a rich set of built-in tools:”
gptme.org · checked Aug 2, 2026
Multi-agent: Comparison table lists multi-agent / concurrent agents support (checkmark), vs. compared tools which lack it.
“Multi-agent ✅ Concurrent agents ❌ ❌ ❌ ”
gptme.org · checked Aug 2, 2026
API access: Provides a Web UI and REST API (via gptme-server).
“🌐 Web UI and REST API”
gptme.org · checked Aug 2, 2026
Open source: Explicitly described as free and open-source.
“Free and open-source.”
gptme.org · checked Aug 2, 2026
Marketplace presence: no public information
gptme.org · checked Aug 2, 2026 · source did not state this

Impact

User base: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Deployment scale: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Target sectors: Targets general knowledge-work use cases: coding, research, automation, documentation, testing, and DevOps.
“gptme is general-purpose but excels at: Coding : Write, refactor, debug code Research : Web browsing, data collection Automation : File management, CI tasks Documentation : Generate docs, summaries Testing : Write tests, run tests, fix failures DevOps : Server management, deployment”
gptme.org · checked Aug 2, 2026
High-risk domains: no public information
gptme.org · checked Aug 2, 2026 · source did not state this
Documented incidents: no public information
gptme.org · checked Aug 2, 2026 · source did not state this