Gratis por defecto
A built-in AI provider is included and pre-configured. Nothing to sign up for, nothing to pay before your first task.
AlphaCode es una herramienta gratuita y de código abierto escrita en Rust. Describe lo que quieres en lenguaje normal: lee tu proyecto, edita archivos y ejecuta tus pruebas. Sin clave de API ni registro.
alphacode, y describe lo que quieres cambiar.
$ alphacode --demo
Explore the demo preview, then install this AI coding assistant and try a real task in your own project. The built-in free AI provider means no API key is required to get started.
$ cat WHY_FREE.md
Most agentic CLIs make you paste in a provider key before they'll do anything. AlphaCode ships with a free model lane already wired in, so the very first thing you run can be a real task against a real repo.
A built-in AI provider is included and pre-configured. Nothing to sign up for, nothing to pay before your first task.
Install, launch alphacode, describe what you need. The free lane answers immediately.
Connect Claude, GPT, Gemini, Copilot, or 50+ others any time with Ctrl+T — no reinstall, same workflow.
Swarm mode, desktop control, 40+ tools, browser automation — all available on the free lane, not a limited trial.
$ alphacode # free model, already configured — no key needed
$ alphacode --help
Reads your codebase, plans steps, picks the best model, edits files, runs tests, and self-reviews before declaring anything done.
Start on the free built-in lane. Add Claude, GPT, Gemini, Copilot, Cursor, Bedrock, OpenRouter, or any OpenAI-compatible endpoint. Switch with Ctrl+T.
Big jobs split into independent pieces, run by multiple agents in parallel, then merged and reviewed automatically.
Uses safeguards for selected destructive operations and prompts for risky actions. Review requests, provider settings, and diffs before accepting changes.
File editing, regex/AST search, shell, browser control, web fetch, memory, scheduling, diagrams, PDFs. No plugin surgery.
Sessions saved to disk. alphacode --resume reopens exactly where you left off — even after a crash.
$ alphacode --how-it-works
Unlike a chat window or autocomplete, an AI coding agent acts on your project. AlphaCode follows a simple loop and stops to ask before risky actions.
Open your project, run alphacode and say what you want in plain English, such as "fix the failing login test".
The language model reads the relevant files, plans the steps, uses its tools to edit code and runs your tests to check the result.
Check the diff, accept or reject changes, and resume any session later with alphacode --resume.
New to the idea? Read our guide: what is an AI coding agent?
$ /swarm "migrate auth to OAuth 2.1"
A planner agent decomposes your goal into independent DAG nodes. Multiple agents execute them concurrently — each with the right model — then results merge and review automatically.
$ alphacode provider list
The built-in free lane needs nothing from you. First-class support for the big providers, OpenAI-compatible for the long tail — anything that speaks HTTP + JSON works.
Don't see yours? Point api_base at any OpenAI-compatible endpoint and you're done.
$ curl … | bash
$ curl -fsSL https://raw.githubusercontent.com/dragonked2/alphacode/main/scripts/install.sh | bashPS> iwr -useb https://raw.githubusercontent.com/dragonked2/alphacode/main/scripts/install.ps1 | iex$ git clone https://github.com/dragonked2/alphacode.git && cd alphacode && cargo build --releaseDespués, verifica:
No step here asks for a key. The free provider is ready the moment install finishes. Prefer to read the script first? View install.sh.
$ ps -o rss -p $(pgrep alphacode)
Compare the published historical RAM snapshots for one and ten concurrent sessions. Results depend on the operating system, build, enabled features, workload, and measurement method.
These are historical snapshots, not a universal performance ranking. Reproduce the same workload and environment before comparing tools. See the benchmark methodology and the safety model for details.
$ alphacode --examples
AlphaCode is an agentic CLI: you describe an outcome, it reads your repository, makes the edits, runs your tests, and reports back. Here are the jobs developers use it for most.
Point AlphaCode at a failing test suite. It reproduces the error, finds the cause with regex and AST search, patches the code and re-runs the tests until they pass.
Rename APIs, upgrade frameworks or move an auth system to OAuth 2.1. Swarm mode splits large migrations into parallel tasks and merges the reviewed result.
Drive a real browser for end-to-end checks, scrape documentation, or operate desktop apps from the same terminal session you use for coding.
Register Model Context Protocol servers to give the agent new tools, such as security scanners or internal APIs. See the MCP guide.
AlphaCode is a single Rust binary with no GUI requirement, so it runs inside tmux, screen, CI runners and remote boxes just as well as on a laptop.
Start on the built-in free provider, then switch to Claude, GPT, Gemini, DeepSeek, Mistral, Ollama or any OpenAI-compatible endpoint without reinstalling.
$ man alphacode | grep -A2 FAQ
No. El proveedor de IA gratuito se configura solo tras la instalación. Más adelante puedes usar Claude, GPT, Gemini o un modelo local con Ollama.
AlphaCode runs on your machine, but it may send task context to the active model provider to generate responses. Which data is sent depends on the task, tools used, and provider configuration. Review provider privacy terms and avoid including secrets in prompts or files shared with a model.
The built-in free provider works out of the box. AlphaCode also supports a range of third-party integrations and OpenAI-compatible endpoints. The available providers may change by release; check the documentation for current setup details.
AlphaCode (también escrito Alpha Code) es un agente de programación con IA gratuito y de código abierto, escrito en Rust, que se ejecuta en tu terminal. Lee tu proyecto, edita archivos y ejecuta pruebas.
No. Alpha Coders y Wallpaper Abyss es un sitio de fondos de pantalla independiente. AlphaCode es una herramienta para desarrolladores y no está vinculada a él.
A planner turns your goal into a task DAG. Workers — each with their own model — execute independent nodes in parallel. A reviewer node merges and verifies the diff.
Yes. Pure terminal. Pipe-friendly. Plays well with tmux, screen, and SSH.
AlphaCode is MIT-licensed, includes a built-in free AI provider, supports multiple model providers, offers swarm workflows, and is written in Rust. The benchmark section provides historical memory measurements for specific test conditions; results vary by version, configuration, and workload.
An AI coding agent is software that uses a large language model to complete programming tasks on your behalf. It reads your codebase, plans the work, edits files, runs commands and tests, and reports back, instead of only suggesting the next line. AlphaCode is a free, open-source AI coding agent that runs in your terminal.
AlphaCode is an open-source, terminal-based AI coding agent in the same category. It is MIT-licensed, written in Rust, includes a free built-in AI provider and can also connect to Claude, GPT, Gemini and other providers you already use. Compare features and privacy terms for your own workflow.
Yes. Ollama and LM Studio are listed among the supported providers, and any OpenAI-compatible endpoint can be used by setting its api_base. This lets you run an AI coding agent against a model on your own machine.
The AlphaCode Discord is where users and maintainers swap workflows, debug together and decide what ships next. Everyone is welcome, from first install to power user.
Únete a DiscordRelease notes, tips and workflows land on X, Facebook and LinkedIn first. Built and maintained by Ali Essam.
Open any project, run alphacode, and describe what you want changed — the free model answers right away.