1# Agent Panel
  2
  3The Agent Panel provides you with a way to interact with LLMs.
  4You can use it for various tasks, such as generating code, asking questions about your code base, and general inquiries such as emails and documentation.
  5
  6To open the Agent Panel, use the `agent: new thread` action in [the Command Palette](./getting-started.md#command-palette) or click the ✨ (sparkles) icon in the status bar.
  7
  8If you're using the Agent Panel for the first time, you'll need to [configure at least one LLM provider](./configuration.md).
  9
 10## Overview {#overview}
 11
 12After you've configured one or more LLM providers, type at the message editor and hit `enter` to submit your prompt.
 13If you need extra room to type, you can expand the message editor with {#kb agent::ExpandMessageEditor}.
 14
 15You should start to see the responses stream in with indications of [which tools](./tools.md) the AI is using to fulfill your prompt.
 16
 17### Editing Messages {#editing-messages}
 18
 19Any message that you send to the AI is editable.
 20You can click on the card that contains your message and re-submit it with an adjusted prompt and/or new pieces of context.
 21
 22### Checkpoints {#checkpoints}
 23
 24Every time the AI performs an edit, you should see a "Restore Checkpoint" button to the top of your message, allowing you to return your codebase to the state it was in prior to that message.
 25
 26The checkpoint button appears even if you interrupt the thread midway through an edit attempt, 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.
 27
 28### Navigating History {#navigating-history}
 29
 30To quickly navigate through recently opened threads, use the {#kb agent::ToggleNavigationMenu} binding, when focused on the panel's editor, or click the hamburger icon button at the top left of the panel to open the dropdown that shows you the six most recent threads.
 31
 32The items in this menu function similarly to tabs, and closing them doesn’t delete the thread; instead, it simply removes them from the recent list.
 33
 34To view all historical conversations, reach for the `View All` option from within the same menu or via the {#kb agent::OpenHistory} binding.
 35
 36### Following the Agent {#following-the-agent}
 37
 38Zed is built with collaboration natively integrated.
 39This approach extends to collaboration with AI as well.
 40To follow the agent reading through your codebase and performing edits, click on the "crosshair" icon button at the bottom left of the panel.
 41
 42### Get Notified {#get-notified}
 43
 44If you send a prompt to the Agent and then move elsewhere, thus putting Zed in the background, you can be notified of whether its response is finished either via:
 45
 46- a visual notification that appears in the top right of your screen
 47- or a sound notification
 48
 49Both notification methods can be used together or individually according to your preference.
 50
 51You can customize their behavior, including turning them off entirely, by using the `agent.notify_when_agent_waiting` and `agent.play_sound_when_agent_done` settings keys.
 52
 53### Reviewing Changes {#reviewing-changes}
 54
 55Once the agent has made changes to your project, the panel will surface which files, and how many of them, have been edited.
 56
 57To 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 multi-buffer tab with all changes.
 58
 59You're able to reject or accept each individual change hunk, or the whole set of changes made by the agent.
 60
 61Edit diffs also appear in individual buffers.
 62So, if your active tab had edits made by the AI, you'll see diffs with the same accept/reject controls as in the multi-buffer.
 63
 64## Adding Context {#adding-context}
 65
 66Although Zed's agent is very efficient at reading through your codebase to autonomously pick up relevant files, directories, and other context, manually adding context is still encouraged as a way to speed up and improve the AI's response quality.
 67
 68If you have a tab open when opening the Agent Panel, that tab appears as a suggested context in form of a dashed button.
 69You can also add other forms of context by either mentioning them with `@` or hitting the `+` icon button.
 70
 71You can even add previous threads as context by mentioning them with `@thread`, or by selecting the "New From Summary" option from the top-right menu to continue a longer conversation, keeping it within the context window.
 72
 73Pasting images as context is also supported by the Agent Panel.
 74
 75### Token Usage {#token-usage}
 76
 77Zed surfaces how many tokens you are consuming for your currently active thread in the panel's toolbar.
 78Depending on how many pieces of context you add, your token consumption can grow rapidly.
 79
 80With that in mind, once you get close to the model's context window, a banner appears below the message editor suggesting to start a new thread with the current one summarized and added as context.
 81You can also do this at any time with an ongoing thread via the "Agent Options" menu on the top right.
 82
 83## Changing Models {#changing-models}
 84
 85After you've configured your LLM providers—either via [a custom API key](./configuration.md#use-your-own-keys) or through [Zed's hosted models](./models.md)—you can switch between them by clicking on the model selector on the message editor or by using the {#kb agent::ToggleModelSelector} keybinding.
 86
 87## Using Tools {#using-tools}
 88
 89The new Agent Panel supports tool calling, which enables agentic editing.
 90Zed comes with [several built-in tools](./tools.md) that allow models to perform tasks such as searching through your codebase, editing files, running commands, and others.
 91
 92You can also extend the set of available tools via [MCP Servers](./mcp.md).
 93
 94### Profiles {#profiles}
 95
 96Profiles act as a way to group tools.
 97Zed offers three built-in profiles and you can create as many custom ones as you want.
 98
 99#### Built-in Profiles {#built-in-profiles}
100
101- `Write`: A profile with tools to allow the LLM to write to your files and run terminal commands. This one essentially has all built-in tools turned on.
102- `Ask`: A profile with read-only tools. Best for asking questions about your code base without the concern of the agent making changes.
103- `Minimal`: A profile with no tools. Best for general conversations with the LLM where no knowledge of your code base is necessary.
104
105You can explore the exact tools enabled in each profile by clicking on the profile selector button > `Configure Profiles…` > the one you want to check out.
106
107#### Custom Profiles {#custom-profiles}
108
109You can create a custom profile via the `Configure Profiles…` option in the profile selector.
110From here, you can choose to `Add New Profile` or fork an existing one with a custom name and your preferred set of tools.
111
112You can also override built-in profiles.
113With a built-in profile selected, in the profile selector, navigate to `Configure Tools`, and select the tools you'd like.
114
115Zed will store this profile in your settings using the same profile name as the default you overrode.
116
117All custom profiles can be edited via the UI or by hand under the `assistant.profiles` key in your `settings.json` file.
118
119### Model Support {#model-support}
120
121Tool calling needs to be individually supported by each model and model provider.
122Therefore, despite the presence of tools, some models may not have the ability to pick them up yet in Zed.
123You should see a "No tools" label if you select a model that falls into this case.
124
125We want to support all of them, though!
126We may prioritize which ones to focus on based on popularity and user feedback, so feel free to help and contribute to fast-track those that don't fit this bill.
127
128All [Zed's hosted models](./models.md) support tool calling out-of-the-box.
129
130### MCP Servers {#mcp-servers}
131
132Similarly to the built-in tools, some models may not support all tools included in a given MCP Server.
133Zed's UI will inform about this via a warning icon that appears close to the model selector.
134
135## Text Threads {#text-threads}
136
137["Text threads"](./text-threads.md) present your conversation with the LLM in a different format—as raw text.
138With text threads, you have full control over the conversation data.
139You can remove and edit responses from the LLM, swap roles, and include more context earlier in the conversation.
140
141For users who have been with us for some time, you'll notice that text threads are our original assistant panel—users love it for the control it offers.
142We do not plan to deprecate text threads, but it should be noted that if you want the AI to write to your code base autonomously, that's only available in the newer, and now default, "Threads".
143
144### Text Thread History {#text-thread-history}
145
146Content from text thread are saved to your file system.
147Visit [the dedicated docs](./text-threads.md#history) for more info.
148
149## Errors and Debugging {#errors-and-debugging}
150
151In case of any error or strange LLM response behavior, the best way to help the Zed team debug is by reaching for the `agent: open thread as markdown` action and attaching that data as part of your issue on GitHub.
152
153This action exposes the entire thread in the form of Markdown and allows for deeper understanding of what each tool call was doing.
154
155You 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.
156
157## Feedback {#feedback}
158
159Every change we make to Zed's system prompt and tool set, needs to be backed by an eval with good scores.
160
161Every time the LLM performs a weird change or investigates a certain topic in your codebase completely incorrectly, it's an indication that there's an improvement opportunity.
162
163> Note that rating responses will send your data related to that response to Zed's servers.
164> 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.
165> **_If you don't want data persisted on Zed's servers, don't rate_**. We will not collect data for improving our Agentic offering without you explicitly rating responses.
166
167The best way you can help influence the next change to Zed's system prompt and tools is by rating the LLM's response via the thumbs up/down buttons at the end of every response.
168In case of a thumbs down, a new text area will show up where you can add more specifics about what happened.
169
170You can provide feedback on the thread at any point after the agent responds, and multiple times within the same thread.