FAQ

Is AlbertCode open source?

No. AlbertCode SWE Agent and mini-AlbertCode SWE Agent are proprietary applications. Their public repositories hold the installers, documentation, examples and benchmark methodology — code under Apache-2.0, documentation under CC BY 4.0 — and welcome contributions. The engine (planning, verification, review), model prompts and orchestration aren't published. See Open Source.

Does it cost anything?

Installing is free and needs no account. AlbertCode doesn't charge for model usage: you use your own model provider account, or a local model at no cost. The repository notes that terms for AlbertCode itself will be published with its first public release.

Does it send my code anywhere?

Only to the model provider you choose — the model needs to read the code to work on it — and to MCP servers you add. There is no telemetry and no account. With a local model, your code never leaves your machine. See each repository's PRIVACY.md.

Which models can I use?

Hosted providers with an API key, Kimi, Hugging Face, OpenRouter, any compatible API service, or a local model server such as Ollama, LM Studio, vLLM or llama.cpp. The model must support tool calling. See Configuration.

Can it change my files without asking?

Not in Governed mode, the default: you approve the plan, then the diff. Direct mode asks before each write, unless you explicitly choose auto. Scripts can answer gates for you with --approve, which you name explicitly.

What happens if I edit a file while it's working?

In Governed mode the work happens in an isolated copy, so your edits are safe. AlbertCode SWE Agent refuses acceptance if you edited a file the change also touches, rather than overwriting your edit.

Does it work offline?

With a local model, yes, once installed. Installing needs the network, and so do hosted models.

Can I run it on this website, or in the cloud?

No. Both agents run on your own computer. The run planner prepares commands and reads results in your browser; it never executes anything. AlbertCode's own browser workspace runs on your computer: albertcode web. Both agents can talk to a service your organisation runs (albertcode --url, albert-mini attach), but no hosted service is offered. See Agent execution.

Is there a Docker image?

No official container image is published. The installers set the agent up for your user account.

What is mini-AlbertCode?

A second agent on the AlbertCode engine, with a full-screen terminal and a browser workspace. AlbertCode SWE Agent is the main one: it adds a VS Code panel and documents the repository contract, the evidence ledger, and the requirements, architecture and browser-test pages. mini-AlbertCode updates itself with albert-mini update. The two install side by side with nothing shared. Compare them.

Where are the benchmark results?

AlbertCode SWE Agent's repository publishes one: SWE-bench Pro HARD-51, 39 of 51 resolved (14 of 22 on tasks held out from development), self-reported with its method, per-task outcomes and limitations. It hasn't been independently verified. See the report.

How do I report a security problem?

Privately, through the repository's Security tab → Report a vulnerability. Please don't open a public issue. Reports are acknowledged within five working days.

Can I contribute?

Yes: documentation, examples, the installer, integrations, tutorials and benchmark reproduction. See Open Source.