AI Frameworks · gptme

gptme

gptme is an open-source AI agent that runs in your terminal, equipped with tools to write code, execute shell commands, and browse the web. It works with multiple LLM providers (OpenAI, Anthropic, local models) and supports autonomous agent creation and extensions via plugins.

Talk to DEV.co Discuss Implementation

Source: GitHub — github.com/gptme/gptme

4.4k

GitHub stars

395

Forks

Python

Primary language

MIT

License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

Field	Value
Repository	gptme/gptme
Owner	gptme
Primary language	Python
License	MIT — OSI-approved
Stars	4.4k
Forks	395
Open issues	5
Latest release	v0.31.0 (2025-12-15)
Last updated	2026-07-08
Source	https://github.com/gptme/gptme

What gptme is

Python-based CLI agent framework supporting multiple LLM backends (OpenAI, Anthropic, Google, DeepSeek, llama.cpp), with integrated tools for code generation, shell execution, web browsing, vision, and MCP/ACP protocol support. Includes persistence, plugin architecture, and autonomous agent templates.

Quickstart

Get the gptme source

Clone the repository and explore it locally.

terminalbash

1git clone https://github.com/gptme/gptme.git2cd gptme3# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Terminal-based code generation and debugging

Automates writing, testing, and fixing code across multiple languages within shell sessions, tmux, or SSH environments. Useful for rapid iteration on scripts, debugging, and autonomous code fixes.

Local-first autonomous agents

Build persistent, self-contained agents using the gptme-agent-template with run loops, GitHub monitoring, and task execution. Runs on headless servers and CI pipelines without external cloud dependencies.

Knowledge-work assistance (research, web scraping, content generation)

Leverages web browsing and vision capabilities to answer questions, extract information from URLs, and assist with writing and analysis tasks in a privacy-respecting, provider-agnostic manner.

Implementation considerations

Requires Python environment and PyPI installation; llama.cpp local mode requires additional setup and local model files.
API keys/credentials for chosen LLM provider (OpenAI, Anthropic, etc.) must be configured; local mode avoids this but requires compatible hardware.
Tool execution (shell, code) runs in the user's environment with broad permissions; security posture depends on agent configuration and guardrails usage.
Persistent agents (via gptme-agent-template) need state storage and may require CI/GitHub integration setup for autonomous operation.
Plugin system (v0.30.0+) and MCP/ACP integrations add extensibility but require Python knowledge to develop custom tools.

When to avoid it — and what to weigh

Require enterprise SLA or commercial support — gptme is community-driven open-source with no formal enterprise support, guaranteed SLA, or commercial backing. Suitable for self-supported teams only.
Need visual GUI-first interaction (not terminal-comfortable) — Primary interface is terminal-based CLI. While web UI exists (gptme-webui, chat.gptme.org), the project is optimized for terminal environments and may frustrate non-CLI-native users.
Strict compliance or audit requirements — Data handling, audit logs, and compliance certifications are not documented. Use with external LLM providers means data flows to those services unless fully local (llama.cpp) is used.
Require guaranteed backwards compatibility — Active development with frequent releases (v0.31.0 Dec 2025) means breaking changes are possible. Plugin and integration APIs may shift across versions.

License & commercial use

MIT License (permissive open-source). Permits commercial use, modification, and distribution with attribution and no warranty. No proprietary restrictions.

MIT allows commercial use without restriction. However, gptme itself provides no indemnification, warranty, or support. Ensure your LLM provider (OpenAI, Anthropic, etc.) permits your commercial use case. No commercial support or SLA from gptme project itself.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

Signal	Assessment
Maintenance	Active
Documentation	Strong
License clarity	Clear
Deployment complexity	Moderate
DEV.co fit	Good
Assessment confidence	High

Security considerations

gptme executes arbitrary shell commands and code in the user's environment. Guardrails feature exists (mentioned in README) but specific implementation details not provided. Data sent to external LLM providers unless local llama.cpp mode is used. No security audit or CVE history provided. Token awareness and cost tracking (v0.31.0) help prevent runaway spending but do not eliminate attack surface. Recommend running in isolated environments and reviewing guardrails configuration before autonomous operation.

Alternatives to consider

Claude Code (Anthropic)

Cloud-hosted coding agent; tightly integrated with Claude model; managed infrastructure and support. Trade-off: less local control, vendor lock-in, requires subscription.

Cursor (IDE-based)

IDE-native AI assistant with strong editor integration; popular in developer workflows. Trade-off: not terminal-first, requires IDE; different UX model.

Warp (AI shell)

Terminal-native AI-assisted shell with command prediction and history. Trade-off: shell replacement, not full agent framework; focused narrower use case.

Software development agency

Build on gptme with DEV.co software developers

gptme is a powerful, actively maintained open-source agent framework ideal for teams comfortable with terminal environments and self-support. Assess provider lock-in, guardrails, and compliance requirements before production use. Contact our engineering team to review architecture fit and integration strategy.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

shell_gpt openai-agents-python SWE-agent

Related on DEV.co

Explore the category and the services that help you build with it.

AI Frameworks AI Application Development API Development DevOps Talk to DEV.co

gptme FAQ

Can I run gptme fully locally without sending data to OpenAI or Anthropic?

Yes, via llama.cpp integration. You can use fully local open models. However, llama.cpp setup and model selection requires additional knowledge and hardware resources.

Is gptme suitable for production autonomous agents?

Possibly, but with caution. The gptme-agent-template provides a framework for persistent agents. Ensure guardrails, monitoring, and error handling suit your risk profile. No formal production SLA or support.

What happens if my LLM provider (OpenAI/Anthropic) changes pricing or policies?

gptme is provider-agnostic, so you can switch providers. However, agent behavior and token costs may vary. Cost tracking (v0.31.0) helps monitor spend.

Can I extend gptme with custom tools?

Yes, via the plugin system (v0.30.0+) and MCP/ACP protocols. Requires Python knowledge. Community plugins exist in gptme-contrib as examples.

Custom software development services

Adopting gptme is usually one piece of a larger software development effort. As a software development agency, DEV.co provides software development services and web development expertise — pairing senior software developers and web developers with your team to design, build, and operate ai frameworks software in production.

Evaluate gptme for Your AI Automation Needs

Talk to DEV.co Schedule a Consultation