1. Previously, checkpoints only appeared when an agent's edit happened
immediately after a user message. This is rare (agent usually collects
some context first), so they were almost never shown. This is now fixed.
2. After this change, a checkpoint is created after every edit
operation. So when the agent edits files five times in a single dialog
turn, we will now display five checkpoints.
As a bonus, it's now possible to undo only a part of a long agent
response.
Closes#36092, #32917
Release Notes:
- Create agent checkpoints more frequently (before every edit)
This was causing text thread summarization to be counted as a usage of 1
prompt
Release Notes:
- Fixed bug with agent text threads (not chat threads) counting
summarization as a usage of 1 prompt.
Co-authored-by: Oleksiy <oleksiy@zed.dev>
This pull request should be idempotent, but lays the groundwork for
avoiding to connect to collab in order to interact with AI features
provided by Zed.
Release Notes:
- N/A
---------
Co-authored-by: Marshall Bowers <git@maxdeviant.com>
Co-authored-by: Richard Feldman <oss@rtfeldman.com>
This PR updates the Agent panel to work with the `CloudUserStore`
instead of the `UserStore`, reducing its reliance on being connected to
Collab to function.
Release Notes:
- N/A
---------
Co-authored-by: Richard Feldman <oss@rtfeldman.com>
Now the edit tool can access files outside the current project (just
like the terminal tool can), but it's behind a prompt (unlike other edit
tool actions).
Release Notes:
- The edit tool can now access files outside the current project, but
only if the user grants it permission to.
Someone encountered this in production, which should not happen:
<img width="1266" height="623" alt="Screenshot 2025-07-24 at 10 38
40 AM"
src="https://github.com/user-attachments/assets/40f3f977-5110-4808-a456-7e708d953b3b"
/>
This moves certain errors into the category of "never retry" and reduces
the number of retries for some others. Also it adds some diagnostic
logging for retry policy.
It's not a complete fix for the above, because the underlying issues is
that the server is sending a HTTP 403 response and although we were
already treating 403s as "do not retry" it was deciding to retry with 2
attempts anyway. So further debugging is needed to figure out why it
wasn't going down the 403 branch by the time the request got here.
Release Notes:
- N/A
TODO
- [x] OpenAI Compatible API Icon
- [x] Docs
- [x] Link to docs in OpenAI provider section about configuring OpenAI
API compatible providers
Closes#33992
Related to #30010
Release Notes:
- agent: Add support for adding multiple OpenAI API compatible providers
---------
Co-authored-by: MrSubidubi <dev@bahn.sh>
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
We no longer rely on the `author` field to tell if a change was made by
the user or the agent. The `author` can be set to `User` in many
situations that are not really user-made edits, such as saving a file,
accepting a change, auto-formatting, and more. I started tracking and
fixing some of these cases, but found that inspecting changes in
`diff_base` is a more reliable method.
Also, we no longer show empty diffs. For example, if the user adds a
line and then removes the same line, the final diff is empty, even
though the buffer is marked as user-changed. Now we won't show such
edit.
There are still some issues to address:
- When a user edits within an unaccepted agent-written block, this
change becomes a part of the agent's edit. Rejecting this block will
lose user edits. It won't be displayed in project notifications, either.
- Accepting an agent block counts as a user-made edit.
- Agent start to call `project_notifications` tool after seeing enough
auto-calls.
Release Notes:
- N/A
Now we explicitly carve out exceptions for which HTTP responses we do
*not* retry for, and retry at least once on all others.
Release Notes:
- The Agent panel now automatically retries failed requests under more
circumstances.
Now we only auto-retry if burn mode is enabled. We also show a "Retry"
button (so you don't have to type "continue") if you think that's the
right remedy, and additionally we show a "Retry and Enable Burn Mode"
button if you don't have it enabled.
<img width="484" height="260" alt="Screenshot 2025-07-17 at 6 25 27 PM"
src="https://github.com/user-attachments/assets/dc5bf1f6-8b11-4041-87aa-4f37c95ea9f0"
/>
<img width="478" height="307" alt="Screenshot 2025-07-17 at 6 22 36 PM"
src="https://github.com/user-attachments/assets/1ed6578a-1696-449d-96d1-e447d11959fa"
/>
Release Notes:
- Only auto-retry Agent requests when Burn Mode is enabled
This tool needs more polishing before being generally available.
Release Notes:
- agent: Disabled `project_notifications` tool by default for the time
being
This change improves user/agent collaborative editing.
When the user edits files that are used by the agent, the
`project_notification` tool now pushes *diffs* of the changes, not just
file names. This helps the agent to stay up to date without needing to
re-read files.
Release Notes:
- Improved user/agent collaborative editing: agent now receives diffs of
user edits
In #33275 I was very conservative about when to retry when there are
errors in language completions in the Agent panel.
Now we retry in more scenarios (e.g. HTTP 5xx and 4xx errors that aren't
in the specific list of ones that we handle differently, such as 429s),
and also we show a notification if the thread halts for any reason.
<img width="441" height="68" alt="Screenshot 2025-07-15 at 12 51 30 PM"
src="https://github.com/user-attachments/assets/433775d0-a8b2-403d-9427-1e296d164980"
/>
<img width="482" height="322" alt="Screenshot 2025-07-15 at 12 44 15 PM"
src="https://github.com/user-attachments/assets/5a508224-0fe0-4d34-9768-25d95873eab8"
/>
Release Notes:
- Automatic retry for more Agent errors
- Whenever the Agent stops, play a sound (if configured) and show a
notification (if configured) if the Zed window was in the background.
This introduces a new field `thinking_allowed` on `LanguageModelRequest`
which lets us control whether thinking should be enabled if the model
supports it.
We permit thinking in the Inline Assistant, Edit File tool and the Git
Commit message generator, this should make generation faster when using
a thinking model, e.g. `claude-sonnet-4-thinking`
Release Notes:
- N/A
Previously, we sent notifications repeatedly until the agent read a
file, which was often inefficient. With this change, we now send a
notification only once (unless the files are modified again, in which
case we'll send another notification).
Release Notes:
- N/A
This commit introduces the `project_notifications` tool, which
proactively pushes notifications to the agent.
Unlike other tools, `Thread` automatically invokes this tool on every
turn, even when the LLM doesn't ask for it. When notifications are
available, the tool use and results are inserted into the thread,
simulating an LLM tool call.
As with other tools, users can disable `project_notifications` in
Profiles if they do not want them.
Currently, the tool only notifies users about stale files: that is,
files that have been edited by the user while the agent is also working
on them. In the future, notifications may be expanded to include
compiler diagnostics, long-running processes, and more.
Release Notes:
- Added `project_notifications` tool
Closes#33792
Follow up to #33237 - Turns out my fix for this was not correct
Release Notes:
- agent: Fixed an issue where tools would not work when two MCP servers
provided a tool with the same name
* Updates to `zed_llm_client-0.8.5` which adds support for `retry_after`
when anthropic provides it.
* Distinguishes upstream provider errors and rate limits from errors
that originate from zed's servers
* Moves `LanguageModelCompletionError::BadInputJson` to
`LanguageModelCompletionEvent::ToolUseJsonParseError`. While arguably
this is an error case, the logic in thread is cleaner with this move.
There is also precedent for inclusion of errors in the event type -
`CompletionRequestStatus::Failed` is how cloud errors arrive.
* Updates `PROVIDER_ID` / `PROVIDER_NAME` constants to use proper types
instead of `&str`, since they can be constructed in a const fashion.
* Removes use of `CLIENT_SUPPORTS_EXA_WEB_SEARCH_PROVIDER_HEADER_NAME`
as the server no longer reads this header and just defaults to that
behavior.
Release notes for this is covered by #33275
Release Notes:
- N/A
---------
Co-authored-by: Richard Feldman <oss@rtfeldman.com>
Co-authored-by: Richard <richard@zed.dev>
Seeing this come up in our server logs when sending requests to
Anthropic: `final assistant content cannot end with trailing
whitespace`.
Release Notes:
- agent: Fixed an issue where Anthropic requests would sometimes fail
because of malformed assistant messages
<img width="484" alt="Screenshot 2025-06-25 at 2 26 16 PM"
src="https://github.com/user-attachments/assets/340f15d7-b115-4895-bae8-b12a915bfda1"
/>
<img width="460" alt="Screenshot 2025-06-25 at 2 26 08 PM"
src="https://github.com/user-attachments/assets/6e587a38-d542-405f-809f-402e87520538"
/>
Now we:
* Automatically retry up to 3 times on upstream Overloaded or 500 errors
(currently for Anthropic only; will add others in future PRs)
* Also automatically retry on rate limit errors (using the provided
duration to wait, if we were given one)
* Give you a notification if you don't have Zed open and we stopped the
thread because of an error
Still todo in future PRs:
* Update collab to report Overloaded and 500 errors differently if
collab itself is passing through an upstream error vs not (currently we
report these as "Zed's API is overloaded" when actually it's the
upstream one!)
* Updating providers other than Anthropic to categorize their errors so
that they benefit from this
* Expanding graceful error handling/retry to other things besides
Overloaded and 500 errors (e.g. connection reset)
Release Notes:
- Automatically retry in Agent Panel instead of erroring out when an
upstream AI API is overloaded or 500s
- Show a notification when an Agent thread errors out and Zed is not the
active window
While working on retries, I discovered some opportunities to reduce
cloning of message segments. These segments have full `String`s (not
`SharedString`s), so cloning them means copying cloning all the bytes of
all the strings in the message, which would be nice to avoid!
Release Notes:
- N/A
The agent now checks if a tool is enabled in the current profile before
calling it. Previously, the agent could still call disabled tools, which
commonly happened after switching profiles in the middle of a thread.
Release Notes:
- Fixed a bug where the agent could use disabled tools sometimes
cc @osyvokon
We were seeing a bunch of errors in our backend when people were using
Claude models with thinking enabled.
In the logs we would see
> an error occurred while interacting with the Anthropic API:
invalid_request_error: messages.x.content.0.type: Expected `thinking` or
`redacted_thinking`, but found `text`. When `thinking` is enabled, a
final `assistant` message must start with a thinking block (preceeding
the lastmost set of `tool_use` and `tool_result` blocks). We recommend
you include thinking blocks from previous turns. To avoid this
requirement, disable `thinking`. Please consult our documentation at
https://docs.anthropic.com/en/docs/build-with-claude/extended-thinking
However, this issue did not occur frequently and was not easily
reproducible. Turns out it was triggered by us not correctly handling
[Redacted Thinking
Blocks](https://docs.anthropic.com/en/docs/build-with-claude/extended-thinking#thinking-redaction).
I could constantly reproduce this issue by including this magic string:
`ANTHROPIC_MAGIC_STRING_TRIGGER_REDACTED_THINKING_46C9A13E193C177646C7398A98432ECCCE4C1253D5E2D82641AC0E52CC2876CB
` in the request, which forces `claude-3-7-sonnet` to emit redacted
thinking blocks (confusingly the magic string does not seem to be
working for `claude-sonnet-4`). As soon as we hit a tool call Anthropic
would return an error.
Thanks to @osyvokon for pointing me in the right direction 😄!
Release Notes:
- agent: Fixed an issue where Anthropic models would sometimes return an
error when thinking was enabled
This PR moves the UI-dependent logic in the `agent` crate into its own
crate, `agent_ui`. The remaining `agent` crate no longer depends on
`editor`, `picker`, `ui`, `workspace`, etc.
This has compile time benefits, but the main motivation is to isolate
our core agentic logic, so that we can make agents more
pluggable/configurable.
Release Notes:
- N/A
This PR is in preparation for doing automatic retries for certain
errors, e.g. Overloaded. It doesn't change behavior yet (aside from some
granularity of error messages shown to the user), but rather mostly
changes some error handling to be exhaustive enum matches instead of
`anyhow` downcasts, and leaves some comments for where the behavior
change will be in a future PR.
Release Notes:
- N/A
Closes#31903
Release Notes:
- agent: Fix an issue where an error would occur when MCP servers
specified tools with the same name
---------
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
Having `Thread::last_usage` as an override of the initially fetched
usage could cause the initial usage to be displayed when the current
thread is empty or in text threads. Fix is to just store last usage info
in `UserStore` and not have these overrides
Release Notes:
- Agent: Fixed request usage display to always include the most recently
known usage - there were some cases where it would show the initially
requested usage.
Removing it for two reasons:
1. We need a better implementation that doesn't hurt caching and doesn't
distracts the agent too much (see
https://github.com/zed-industries/zed/pull/32876 for more context)
2. Current insertion point of notifications doesn't play well with
Claude Thinking models (see
https://github.com/zed-industries/zed/issues/33000#issuecomment-2991709484)
I think we should get this code back in a form of a tool. But for now,
I'm dropping it to resolve recent issues.
Closes#33000
Release Notes:
- N/A
Previously we were using a mix of `u32` and `usize`, e.g. `max_tokens:
usize, max_output_tokens: Option<u32>` in the same `struct`.
Although [tiktoken](https://github.com/openai/tiktoken) uses `usize`,
token counts should be consistent across targets (e.g. the same model
doesn't suddenly get a smaller context window if you're compiling for
wasm32), and these token counts could end up getting serialized using a
binary protocol, so `usize` is not the right choice for token counts.
I chose to standardize on `u64` over `u32` because we don't store many
of them (so the extra size should be insignificant) and future models
may exceed `u32::MAX` tokens.
Release Notes:
- N/A
When the user edits one of the tracked files, we used to notify the
agent by inserting a user message at the end of the thread. This was
causing a few problems:
- The agent would stop doing its work and start reading changed files
- The agent would write something like, "Thank you for letting me know
about these changed files."
This fix contains two parts:
1. Changing the prompt to indicate this is a service message
2. Moving the message higher in the conversation thread
This works, but it slightly hurts caching.
We may consider making these notification messages stick in history,
trading context tokens count for the cache.
This might be related to #30906
Release Notes:
- N/A
---------
Co-authored-by: Marshall Bowers <git@maxdeviant.com>
Bubbles up rate limit information so that we can retry after a certain
duration if needed higher up in the stack.
Also caps the number of concurrent evals running at once to also help.
Release Notes:
- N/A
This allows storing the profile per thread, as well as moving the logic
of which tools are enabled or not to the profile itself.
This makes it much easier to switch between profiles, means there is
less global state being changed on every profile change.
Release Notes:
- agent panel: allow saving the profile per thread
---------
Co-authored-by: Ben Kunkle <ben.kunkle@gmail.com>
Co-authored-by: Cole Miller <m@cole-miller.net>
Release Notes:
- agent: Fixed a panic when re-editing old messages
---------
Co-authored-by: Cole Miller <m@cole-miller.net>
Co-authored-by: Cole Miller <cole@zed.dev>
This PR introduces the "Reject All" and "Accept All" buttons in the
panel's edit bar, which appears as soon as the agent starts editing a
file. I'm also adding here a new method to the thread called
`has_pending_edit_tool_uses`, which is a more specific way of knowing,
in comparison to the `is_generating` method, whether or not the
reject/accept all actions can be triggered.
Previously, without this new method, you'd be waiting for the whole
generation to end (e.g., the agent would be generating markdown with
things like change summary) to be able to click those buttons, when the
edit was already there, ready for you. It always felt like waiting for
the whole thing was unnecessary when you really wanted to just wait for
the _edits_ to be done, as so to avoid any potential conflicting state.
<img
src="https://github.com/user-attachments/assets/0927f3a6-c9ee-46ae-8f7b-97157d39a7b5"
width="500"/>
---
Release Notes:
- agent: Added ability to reject and accept all changes from the agent
panel.
---------
Co-authored-by: Agus Zubiaga <hi@aguz.me>
When reaching the consecutive tool call limit, the agent gets blocked
and without a notification, you wouldn't know that. This PR adds the
ability to be notified when that happens, and you can use either sound
_and_ toast, or just one of them.
Release Notes:
- agent: Added support for getting notified (via toast and/or sound)
when reaching the consecutive tool call limit.
