1# Agent Settings
2
3Learn about all the settings you can customize in Zed's Agent Panel.
4
5## Model Settings {#model-settings}
6
7### Default Model {#default-model}
8
9If you're using [Zed's hosted LLM service](./plans-and-usage.md), it sets `claude-sonnet-4` as the default model.
10But if you're not subscribed to it or simply just want to change it, you can do it so either via the model dropdown in the Agent Panel's bottom-right corner or by manually editing the `default_model` object in your settings:
11
12```json
13{
14 "agent": {
15 "default_model": {
16 "provider": "zed.dev",
17 "model": "gpt-4o"
18 }
19 }
20}
21```
22
23### Feature-specific Models {#feature-specific-models}
24
25Assign distinct and specific models for the following AI-powered features in Zed:
26
27- Thread summary model: Used for generating thread summaries
28- Inline assistant model: Used for the inline assistant feature
29- Commit message model: Used for generating Git commit messages
30
31```json
32{
33 "agent": {
34 "default_model": {
35 "provider": "zed.dev",
36 "model": "claude-sonnet-4"
37 },
38 "inline_assistant_model": {
39 "provider": "anthropic",
40 "model": "claude-3-5-sonnet"
41 },
42 "commit_message_model": {
43 "provider": "openai",
44 "model": "gpt-4o-mini"
45 },
46 "thread_summary_model": {
47 "provider": "google",
48 "model": "gemini-2.0-flash"
49 }
50 }
51}
52```
53
54> If a custom model isn't set for one of these features, they automatically fall back to using the default model.
55
56### Alternative Models for Inline Assists {#alternative-assists}
57
58The Inline Assist feature in particular has the capacity to perform multiple generations in parallel using different models.
59That is possible by assigning more than one model to it, taking the configuration shown above one step further.
60
61When configured, the inline assist UI will surface controls to cycle between the outputs generated by each model.
62
63The models you specify here are always used in _addition_ to your [default model](#default-model).
64
65For example, the following configuration will generate two outputs for every assist.
66One with Claude Sonnet 4 (the default model), and one with GPT-4o.
67
68```json
69{
70 "agent": {
71 "default_model": {
72 "provider": "zed.dev",
73 "model": "claude-sonnet-4"
74 },
75 "inline_alternatives": [
76 {
77 "provider": "zed.dev",
78 "model": "gpt-4o"
79 }
80 ]
81 }
82}
83```
84
85### Model Temperature
86
87Specify a custom temperature for a provider and/or model:
88
89```json
90"model_parameters": [
91 // To set parameters for all requests to OpenAI models:
92 {
93 "provider": "openai",
94 "temperature": 0.5
95 },
96 // To set parameters for all requests in general:
97 {
98 "temperature": 0
99 },
100 // To set parameters for a specific provider and model:
101 {
102 "provider": "zed.dev",
103 "model": "claude-sonnet-4",
104 "temperature": 1.0
105 }
106],
107```
108
109## Agent Panel Settings {#agent-panel-settings}
110
111Note that some of these settings are also surfaced in the Agent Panel's settings UI, which you can access either via the `agent: open settings` action or by the dropdown menu on the top-right corner of the panel.
112
113### Default View
114
115Use the `default_view` setting to change the default view of the Agent Panel.
116You can choose between `thread` (the default) and `text_thread`:
117
118```json
119{
120 "agent": {
121 "default_view": "text_thread"
122 }
123}
124```
125
126### Auto-run Commands
127
128Control whether you want to allow the agent to run commands without asking you for permission.
129The default value is `false`.
130
131```json
132{
133 "agent": {
134 "always_allow_tool_actions": true
135 }
136}
137```
138
139> This setting is available via the Agent Panel's settings UI.
140
141### Single-file Review
142
143Control whether you want to see review actions (accept & reject) in single buffers after the agent is done performing edits.
144The default value is `false`.
145
146```json
147{
148 "agent": {
149 "single_file_review": true
150 }
151}
152```
153
154When set to false, these controls are only available in the multibuffer review tab.
155
156> This setting is available via the Agent Panel's settings UI.
157
158### Sound Notification
159
160Control whether you want to hear a notification sound when the agent is done generating changes or needs your input.
161The default value is `false`.
162
163```json
164{
165 "agent": {
166 "play_sound_when_agent_done": true
167 }
168}
169```
170
171> This setting is available via the Agent Panel's settings UI.
172
173### Message Editor Size
174
175Use the `message_editor_min_lines` setting to control minimum number of lines of height the agent message editor should have.
176It is set to `4` by default, and the max number of lines is always double of the minimum.
177
178```json
179{
180 "agent": {
181 "message_editor_min_lines": 4
182 }
183}
184```
185
186> This setting is currently available only in Preview.
187
188### Modifier to Send
189
190Make a modifier (`cmd` on macOS, `ctrl` on Linux) required to send messages.
191This is encouraged for more thoughtful prompt crafting.
192The default value is `false`.
193
194```json
195{
196 "agent": {
197 "use_modifier_to_send": true
198 }
199}
200```
201
202> This setting is available via the Agent Panel's settings UI.
203
204### Edit Card
205
206Use the `expand_edit_card` setting to control whether edit cards show the full diff in the Agent Panel.
207It is set to `true` by default, but if set to false, the card's height is capped to a certain number of lines, requiring a click to be expanded.
208
209```json
210{
211 "agent": {
212 "expand_edit_card": false
213 }
214}
215```
216
217### Terminal Card
218
219Use the `expand_terminal_card` setting to control whether terminal cards show the command output in the Agent Panel.
220It is set to `true` by default, but if set to false, the card will be fully collapsed even while the command is running, requiring a click to be expanded.
221
222```json
223{
224 "agent": {
225 "expand_terminal_card": false
226 }
227}
228```
229
230### Feedback Controls
231
232Control whether you want to see the thumbs up/down buttons to give Zed feedback about the agent's performance.
233The default value is `true`.
234
235```json
236{
237 "agent": {
238 "enable_feedback": false
239 }
240}
241```