Zed
Fast collaborative code editor with AI agentic editing, inline assistance, edit prediction, and configurable model access.
Zed
Explore features, practical uses and pricing below.
Zed is a code editor designed around speed, collaboration, and integrated AI work. Its AI features include agentic editing and other assistance within the editor, while model access can come through hosted options, provider connections, or supported local configurations. This makes it useful for developers who want to keep code navigation and AI work in the same environment.
Features and what they are useful for
- Edit and navigate code in a desktop editor built for responsive interaction.
- Use agentic workflows that can inspect files, propose changes, and work with tools.
- Apply supported inline editing or edit prediction for smaller code changes.
- Configure model access and supported external-agent integrations according to the current documentation.
Who it suits
Zed suits developers who value a lightweight editing environment and want to try AI without treating one model provider as inseparable from the editor. It is particularly useful for comparing manual editing with an agent-assisted workflow on the same project. Collaboration features also matter for teams that occasionally work together in a shared editing session.
A practical workflow
Open a familiar project and first test ordinary navigation, language support, and build commands. Ask the AI to make one small change with a clear acceptance criterion, then inspect the diff and run the relevant check. Compare the agent's behavior with an inline edit for a smaller request. Configure project instructions only after identifying which recurring conventions actually need to be supplied.
Limitations and things to check
Editor responsiveness and model capability are separate concerns. A fast interface does not make every generated change correct, and tool-using agents may need permissions beyond a simple chat response. Check provider data handling, local model capabilities, and external-agent setup. Repository tests and human review remain necessary when AI changes behavior, even if the edit looks plausible in the interface.
Pricing and access
Zed provides its editor through official downloads and offers several model-access routes. Hosted plans, external provider API use, and local inference can have different costs and requirements. Consult current documentation for supported platforms, subscriptions, gateways, and agent connections before assuming one account unlocks every AI feature.
Frequently asked questions
Can I choose a different model provider?
Zed documents multiple model-access routes, including provider and local configurations.
Are external agents the same as the built-in agent?
No. Their integration and model access can differ, so follow the relevant setup documentation.