Agent Panel

The Agent Panel is where you interact with AI agents that can read, write, and run code in your project. It's the core of Zed's AI code editing experience — use it for code generation, refactoring, debugging, documentation, and general questions.

Open it with agent: new thread from the Command Palette or click the ✨ icon in the status bar.

Getting Started

If you're using the Agent Panel for the first time, configure either a model for the Zed Agent or an External Agent.

Use LLM Providers for Zed-hosted models, API access, subscriptions, gateways, and local models.
Use External Agents for ACP-integrated agents.
Use AI Quick Start if you are not sure which path to choose.

Overview

With an LLM provider or External Agent configured, type in the message editor and press enter to submit. Expand the editor with shift-alt-escape|shift-alt-escape if you need more room.

Responses stream in with indicators showing which tools the model is using. The sections below cover what you can do from here.

Note that some Agent Panel features may not be available for every External Agent. Restoring threads from history, checkpoints, token usage display, and similar features depend on the agent integration. Their availability varies depending on the agent.

Creating New Threads

By default, the Agent Panel uses Zed's first-party agent.

Start a new thread with cmd-n|ctrl-n, or open the "New Thread…" menu using the agent selector button on the left (in the empty state) or the + icon in the top-right of the panel toolbar. You can also open that menu with cmd-alt-shift-n|ctrl-alt-shift-n.

From the "New Thread…" menu you can:

Pick Zed Agent or any installed External Agent to start a new thread with that agent.
Choose New From Summary to start a fresh Zed Agent thread seeded with a summary of the current conversation — useful for compacting long threads as you approach the context window limit.
Choose Terminal to open a terminal thread directly in the Agent Panel — see Terminal Threads for details.

agent: new external agent thread creates a new thread with the specified External Agent id.

You can also start a new thread from the Threads Sidebar, scoped to a specific project — see Running Multiple Threads.

Managing Multiple Threads

You can run multiple agent threads at once, each working independently with its own agent, context window, and conversation history. Open the Threads Sidebar with cmd-alt-j|ctrl-alt-j to see all your threads grouped by project. Click any thread to switch to it, or use the thread switcher (ctrl-tab|ctrl-tab) to cycle between recent threads without opening the sidebar.

Threads you're no longer working on can be archived by hovering over them in the sidebar and clicking the archive icon, or selecting them and pressing shift-backspace|shift-backspace. The Thread History holds all your threads across all projects, sorted chronologically, and you can restore them at any time.

If two threads might edit the same files, you can isolate one in a new Git worktree. Use the worktree picker in the title bar to pick which worktree the agent runs in, or create a new one. See Worktree Isolation for details.

For more details on the Threads Sidebar and managing multiple projects, see Parallel Agents.

Editing Messages

Any message that you send to the model is editable. You can click on the card that contains your message and re-submit it with an adjusted prompt and/or new pieces of context.

Queueing Messages

Messages sent while the agent is in the generating state get, by default, queued.

For the Zed Agent, queued messages get sent at the next turn boundary, which is usually between a tool call and a response, whereas for External Agents, the message gets sent at the end of the generation.

You can edit or remove (an individual or all) queued messages. You can also still interrupt the agent immediately if you want by either clicking on the stop button or by clicking the "Send Now" (double-enter) on a queued message.

Checkpoints

Every time the model performs an edit, you should see a "Restore Checkpoint" button at the top of your message, allowing you to return your code base to the state it was in prior to that message.

The checkpoint button appears even if you interrupt the thread midway through an edit, as this is likely a moment when you've identified that the agent is not heading in the right direction and you want to revert back.

Right-click on any agent response in the thread view to access a context menu with the following actions:

Copy Selection: Copies the currently selected text as Markdown (available when text is selected).
Copy This Agent Response: Copies the full text of the agent response you right-clicked on.
Scroll to Top / Scroll to Bottom: Scrolls to the beginning or end of the thread, depending on your current position.
Open Thread as Markdown: Opens the entire thread as a Markdown file in a new tab.

Navigating the Thread

In long conversations, use the scroll arrow buttons at the bottom of the panel to jump to your most recent prompt or to the very beginning of the thread. You can also scroll the thread using arrow keys, Page Up/Down, Home/End, and Shift+Page Up/Down to jump between messages, when the thread pane is focused.

Thread titles

Thread titles are auto-generated based on the content of the conversation. But you can also edit them manually by clicking the title and typing, or regenerate them by clicking the "Regenerate Thread Title" button in the ellipsis menu in the top right of the panel.

Following the Agent

Follow the agent as it reads and edits files by clicking the crosshair icon at the bottom left of the panel. Your editor will jump to each file the agent touches.

You can also hold cmd/ctrl when submitting a message to automatically follow.

Get Notified

If you send a prompt to the Agent and then put Zed in the background, you can choose to be notified when its generation wraps up via:

a visual notification that appears in the top right of your screen
a sound notification

These notifications can be used together or individually, and you can use the agent.notify_when_agent_waiting and agent.play_sound_when_agent_done settings keys to customize that, including turning both off entirely.

Reviewing Changes

Once the agent has made changes to your project, the panel will surface which files, how many of them, and how many lines have been edited.

To see which files specifically have been edited, expand the accordion bar that shows up right above the message editor or click the Review Changes button (shift-ctrl-r|ctrl-shift-r), which opens a special multi-buffer tab with all changes.

You can accept or reject each individual change hunk, or the whole set of changes made by the agent.

Edit diffs can also appear inline in individual files with the same keep/reject hunk controls as the multi-buffer review pane. This temporarily overrides the buffer's git diff while review is active. Enable it by setting agent.single_file_review to true in your settings.

Terminal Threads

The Agent Panel can host Terminal Threads alongside your agent threads. For opening, closing, notifications, terminal titles, and CLI/TUI-specific setup, see Terminal Threads.

Adding Context

The agent can search your codebase to find relevant context, but providing it explicitly improves response quality and reduces latency.

Add context by typing @ in the message editor. You can mention files, directories, symbols, previous threads, skills, instruction files, and diagnostics.

When you paste multi-line code selections copied from a buffer, Zed automatically formats them as @-mentions with the file context. To paste content without this automatic formatting, use cmd-shift-v|ctrl-shift-v to paste raw text directly.

Selection as Context

Additionally, you can also select text in a buffer or terminal and add it as context by using the cmd->|ctrl-> keybinding, running the agent: add selection to thread action, or choosing the "Selection" item in the + menu in the message editor.

Images as Context

It's also possible to attach images in your prompt for providers that support vision models. OpenAI GPT-4o and later, Anthropic Claude 3 and later, Google Gemini 1.5 and 2.0, and Bedrock vision models (Claude 3+, Amazon Nova Pro and Lite, Meta Llama 3.2 Vision, Mistral Pixtral) all support image inputs.

To add an image, you can either search in your project's folder by @-mentioning it, or drag it from your file system directly into the Agent Panel message editor. Copying an image and pasting it is also supported.

Token Usage

Zed surfaces how many tokens you are consuming for your currently active thread near the profile selector in the panel's message editor.

Once you approach the model's context window, a banner appears above the message editor suggesting to start a new thread with the current one summarized and added as context. You can also do this at any time with an ongoing thread via the "Agent Options" menu on the top right, where you'll see a "New from Summary" button, as well as simply @-mentioning a past thread in a new one..

Changing Models

After you've configured your LLM providers—either via API access or through Zed-hosted models—you can switch between their models by clicking on the model selector on the message editor or by using the cmd-alt-/|ctrl-alt-/ keybinding.

The same model can be offered via multiple providers - for example, Claude Sonnet 4.5 is available via Zed Pro, OpenRouter, Anthropic directly, and more. Make sure you've selected the correct model provider for the model you'd like to use, delineated by the logo to the left of the model in the model selector.

Favoriting Models

You can mark specific models as favorites either through the model selector, by clicking on the star icon button that appears as you hover the model, or through your settings via the agent.favorite_models settings key.

Cycle through your favorites with alt-tab|alt-l without opening the model selector.

Using Tools and Profiles

The Agent Panel supports tool calling, which enables agentic editing. Zed includes built-in tools for searching your codebase, editing files, running terminal commands, and more.

Use Agent Profiles to choose which built-in tools and MCP tools are available in a Zed Agent thread. Use Tool Permissions to control whether permission-gated tool calls are allowed, denied, or confirmed.

You can add external tools with MCP Servers.

Model Support

Tool calling needs to be individually supported by each model and model provider. Therefore, despite the presence of built-in tools, some models may not have the ability to pick them up. You should see a "No tools" label if you select a model that falls into this case.

All Zed-hosted models support tool calling out-of-the-box.

MCP Servers

Similarly to the built-in tools, some models may not support all tools included in a given MCP Server. Zed's UI will inform you about this via a warning icon that appears close to the model selector.

Errors and Debugging

If you hit an error or unusual LLM behavior, open the thread as Markdown with agent: open active thread as markdown and attach it to your GitHub issue.

You can also open threads as Markdown by clicking on the file icon button, to the right of the thumbs down button, when focused on the panel's editor.

Feedback

You can rate agent responses to help improve Zed's system prompt and tools.

Warning: Rating an AI response sends the conversation thread to Zed. The conversation thread includes your messages, AI responses, and thread metadata. See Feedback and Training Data and AI Privacy for more information. If you don't want data persisted on Zed's servers, don't rate. We will not collect data for improving Zed's agent experience without you explicitly rating responses.

To help improve Zed's system prompt and tools, rate responses with the thumbs up/down controls at the end of each response. In case of a thumbs down, a new text area will show up where you can add more specifics about what happened.

You can provide feedback on the thread at any point after the agent responds, and multiple times within the same thread.