1---
2title: AI Coding Agent - Zed Agent Panel
3description: Use Zed's AI coding agent to generate, refactor, and debug code with tool calling, checkpoints, and multi-model support.
4---
5
6# Agent Panel
7
8The Agent Panel is where you interact with AI agents that can read, write, and run code in your project.
9It's the core of Zed's AI code editing experience — use it for code generation, refactoring, debugging, documentation, and general questions.
10
11Open it with `agent: new thread` from [the Command Palette](../getting-started.md#command-palette) or click the ✨ icon in the status bar.
12
13## Getting Started {#getting-started}
14
15If you're using the Agent Panel for the first time, you need to have at least one LLM provider or external agent configured.
16You can do that by:
17
181. [subscribing to our Pro plan](https://zed.dev/pricing), so you have access to our hosted models
192. [using your own API keys](./llm-providers.md#use-your-own-keys), either from model providers like Anthropic or model gateways like OpenRouter.
203. using an [external agent](./external-agents.md) like [Gemini CLI](./external-agents.md#gemini-cli) or [Claude Agent](./external-agents.md#claude-agent)
21
22## Overview {#overview}
23
24With an LLM provider or external agent configured, type in the message editor and press `enter` to submit.
25Expand the editor with {#kb agent::ExpandMessageEditor} if you need more room.
26
27Responses stream in with indicators showing [which tools](./tools.md) the model is using.
28The sections below cover what you can do from here.
29
30> Note that for external agents, like [Gemini CLI](./external-agents.md#gemini-cli) or [Claude Agent](./external-agents.md#claude-agent), some of the features outlined below may _not_ be supported—for example, _restoring threads from history_, _checkpoints_, _token usage display_, and others.
31> Their availability varies depending on the agent.
32
33### Creating New Threads {#new-thread}
34
35By default, the Agent Panel uses Zed's first-party agent.
36
37To choose another agent, go to the plus button in the top-right of the Agent Panel and pick one of the [external agents](./external-agents.md) installed out of the box.
38
39### Editing Messages {#editing-messages}
40
41Any message that you send to the model is editable.
42You can click on the card that contains your message and re-submit it with an adjusted prompt and/or new pieces of context.
43
44### Queueing Messages
45
46Messages sent while the agent is in the generating state get, by default, queued.
47
48For 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.
49
50You can edit or remove (an individual or all) queued messages.
51You 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.
52
53### Checkpoints {#checkpoints}
54
55Every 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.
56
57The 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.
58
59### Context Menu {#context-menu}
60
61Right-click on any agent response in the thread view to access a context menu with the following actions:
62
63- **Copy Selection**: Copies the currently selected text as Markdown (available when text is selected).
64- **Copy This Agent Response**: Copies the full text of the agent response you right-clicked on.
65- **Scroll to Top / Scroll to Bottom**: Scrolls to the beginning or end of the thread, depending on your current position.
66- **Open Thread as Markdown**: Opens the entire thread as a Markdown file in a new tab.
67
68### Navigating the Thread {#navigating-the-thread}
69
70In 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.
71
72When focus is in the message editor, you can also use {#kb agent::ScrollOutputPageUp}, {#kb agent::ScrollOutputPageDown}, {#kb agent::ScrollOutputToTop}, {#kb agent::ScrollOutputToBottom}, {#kb agent::ScrollOutputLineUp}, and {#kb agent::ScrollOutputLineDown} to navigate the thread, or {#kb agent::ScrollOutputToPreviousMessage} and {#kb agent::ScrollOutputToNextMessage} to jump between your prompts.
73
74### Thread titles
75
76Thread titles are auto-generated based on the content of the conversation.
77But 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.
78
79### Following the Agent {#following-the-agent}
80
81Follow the agent as it reads and edits files by clicking the crosshair icon at the bottom left of the panel.
82Your editor will jump to each file the agent touches.
83
84You can also hold `cmd`/`ctrl` when submitting a message to automatically follow.
85
86### Get Notified {#get-notified}
87
88If 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:
89
90- a visual notification that appears in the top right of your screen
91- a sound notification
92
93These 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.
94
95### Reviewing Changes {#reviewing-changes}
96
97Once 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.
98
99To 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 ({#kb agent::OpenAgentDiff}), which opens a special multi-buffer tab with all changes.
100
101You can accept or reject each individual change hunk, or the whole set of changes made by the agent.
102
103Edit diffs also appear in singleton buffers.
104If your active tab had edits made by the AI, you'll see diffs with the same accept/reject controls as in the multi-buffer.
105You can turn this off, though, through the `agent.single_file_review` setting.
106
107## Adding Context {#adding-context}
108
109The agent can search your codebase to find relevant context, but providing it explicitly improves response quality and reduces latency.
110
111Add context by typing `@` in the message editor.
112You can mention files, directories, symbols, previous threads, rules files, and diagnostics.
113
114When you paste multi-line code selections copied from a buffer, Zed automatically formats them as @-mentions with the file context.
115To paste content without this automatic formatting, use {#kb agent::PasteRaw} to paste raw text directly.
116
117### Selection as Context
118
119Additionally, you can also select text in a buffer or terminal and add it as context by using the {#kb agent::AddSelectionToThread} keybinding, running the {#action agent::AddSelectionToThread} action, or choosing the "Selection" item in the `+` menu in the message editor.
120
121### Images as Context
122
123It's also possible to attach images in your prompt for providers that support vision models.
124OpenAI 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.
125
126To add an image, you can either search in your project's directory by @-mentioning it, or drag it from your file system directly into the agent panel message editor.
127Copying an image and pasting it is also supported.
128
129## Token Usage {#token-usage}
130
131Zed surfaces how many tokens you are consuming for your currently active thread near the profile selector in the panel's message editor.
132
133Once 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.
134You 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..
135
136## Changing Models {#changing-models}
137
138After you've configured your LLM providers—either via [a custom API key](./llm-providers.md) or through [Zed's hosted models](./models.md)—you can switch between their models by clicking on the model selector on the message editor or by using the {#kb agent::ToggleModelSelector} keybinding.
139
140> 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.
141> 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.
142
143### Favoriting Models
144
145You 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.
146
147Cycle through your favorites with {#kb agent::CycleFavoriteModels} without opening the model selector.
148
149## Using Tools {#using-tools}
150
151The Agent Panel supports tool calling, which enables agentic editing.
152Zed includes several [built-in tools](./tools.md) for searching your codebase, editing files, running terminal commands, and more.
153
154You can also extend the set of available tools via [MCP Servers](./mcp.md).
155
156### Profiles {#profiles}
157
158Profiles act as a way to group tools.
159Zed offers three built-in profiles and you can create as many custom ones as you want.
160
161#### Built-in Profiles {#built-in-profiles}
162
163- `Write`: A profile with tools to allow the LLM to write to your files and run terminal commands.
164 This one essentially has all built-in tools turned on.
165- `Ask`: A profile with read-only tools.
166 Best for asking questions about your code base without the concern of the agent making changes.
167- `Minimal`: A profile with no tools.
168 Best for general conversations with the LLM where no knowledge of your code base is necessary.
169
170You can explore the exact tools enabled in each profile by clicking on the profile selector button > `Configure` button > the one you want to check out.
171
172Alternatively, you can also use either the command palette, by running {#action agent::ManageProfiles}, or the keybinding directly, {#kb agent::ManageProfiles}, to have access to the profile management modal.
173
174Use {#kb agent::CycleModeSelector} to cycle through available profiles without opening the modal.
175
176#### Custom Profiles {#custom-profiles}
177
178You can also create a custom profile through the Agent Profile modal.
179From there, you can choose to `Add New Profile` or fork an existing one with a custom name and your preferred set of tools.
180
181It's also possible to override built-in profiles.
182In the Agent Profile modal, select a built-in profile, navigate to `Configure Tools`, and rearrange the tools you'd like to keep or remove.
183
184Zed will store this profile in your settings using the same profile name as the default you overrode.
185
186All custom profiles can be edited via the UI or by hand under the `agent.profiles` key in your settings file.
187
188To delete a custom profile, open the Agent Profile modal, select the profile you want to remove, and click the delete button.
189
190### Tool Permissions
191
192> **Note:** In Zed v0.224.0 and above, tool approval is controlled by `agent.tool_permissions.default`.
193> In earlier versions, it was controlled by the `agent.always_allow_tool_actions` boolean (default `false`).
194
195Zed's Agent Panel provides the `agent.tool_permissions.default` setting to control tool approval behavior:
196
197- `"confirm"` (default) — Prompts for approval before running any tool action
198- `"allow"` — Auto-approves tool actions without prompting
199- `"deny"` — Blocks all tool actions
200
201When the agent requests permission for an action, the confirmation menu includes options to allow or deny once, plus "Always for <tool>" choices that set a tool-level default.
202When Zed can extract a safe pattern from the input, it also offers pattern-based "Always for ..." choices that add `always_allow`/`always_deny` rules.
203MCP tools only support tool-level defaults.
204
205Even with `"default": "allow"`, per-tool `always_deny` and `always_confirm` patterns are still respected — so you can auto-approve most actions while blocking or gating specific ones.
206
207Learn more about [how tool permissions work](./tool-permissions.md), how to further customize them, and other details.
208
209### Model Support {#model-support}
210
211Tool calling needs to be individually supported by each model and model provider.
212Therefore, despite the presence of built-in tools, some models may not have the ability to pick them up.
213You should see a "No tools" label if you select a model that falls into this case.
214
215All [Zed's hosted models](./models.md) support tool calling out-of-the-box.
216
217### MCP Servers {#mcp-servers}
218
219Similarly to the built-in tools, some models may not support all tools included in a given MCP Server.
220Zed's UI will inform you about this via a warning icon that appears close to the model selector.
221
222## Errors and Debugging {#errors-and-debugging}
223
224If you hit an error or unusual LLM behavior, open the thread as Markdown with `agent: open thread as markdown` and attach it to your GitHub issue.
225
226You 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.
227
228## Feedback {#feedback}
229
230You can rate agent responses to help improve Zed's system prompt and tools.
231
232> Note that rating responses will send your data related to that response to Zed's servers.
233> See [AI Improvement](./ai-improvement.md) and [Privacy and Security](./privacy-and-security.md) for more information about Zed's approach to AI improvement, privacy, and security.
234> **_If you don't want data persisted on Zed's servers, don't rate_**.
235> We will not collect data for improving our Agentic offering without you explicitly rating responses.
236
237To help improve Zed's system prompt and tools, rate responses with the thumbs up/down controls at the end of each response.
238In case of a thumbs down, a new text area will show up where you can add more specifics about what happened.
239
240You can provide feedback on the thread at any point after the agent responds, and multiple times within the same thread.