## Summary
Fixes#47540 - Anthropic Claude models not appearing in GitHub Copilot
Chat model picker.
## Problem
Users reported that Anthropic Claude models (Claude Sonnet 4, Claude
Opus 4, etc.) were not appearing in the model picker when using GitHub
Copilot Chat, even though:
- The GitHub Copilot API returns these models
- The models have `model_picker_enabled: true`
- Users have valid Copilot subscriptions with access to these models
## Root Cause
The issue was in the `ModelSupportedEndpoint` enum deserialization. The
enum only defined two variants:
```rust
pub enum ModelSupportedEndpoint {
#[serde(rename = "/chat/completions")]
ChatCompletions,
#[serde(rename = "/responses")]
Responses,
}
```
Anthropic Claude models use the `/v1/messages` endpoint, which wasn't
defined. When deserializing the API response, serde failed with:
```
Error("unknown variant `/v1/messages`, expected `/chat/completions` or `/responses`")
```
Because the crate uses resilient deserialization via
`deserialize_models_skip_errors()`, the entire Claude model was silently
skipped rather than causing a hard failure. This meant users saw no
error - the models simply didn't appear.
## Solution
### 1. Added `/v1/messages` endpoint variant
```rust
pub enum ModelSupportedEndpoint {
#[serde(rename = "/chat/completions")]
ChatCompletions,
#[serde(rename = "/responses")]
Responses,
#[serde(rename = "/v1/messages")]
Messages, // NEW: Anthropic models use this endpoint
#[serde(other)]
Unknown, // NEW: Future-proofing for unknown endpoints
}
```
### 2. Removed incorrect `dedup_by()` call
The previous code deduplicated models by family:
```rust
.dedup_by(|a, b| a.capabilities.family == b.capabilities.family)
```
This incorrectly filtered out model variants that share the same family
(e.g., `claude-sonnet-4` and `claude-sonnet-4-thinking`). Removed this
call to preserve all model variants.
### 3. Removed unused import
Removed `use itertools::Itertools;` which was only used for the
now-removed `dedup_by()`.
## Changes
| File | Change |
|------|--------|
| `crates/copilot_chat/src/copilot_chat.rs` | Added `Messages` and
`Unknown` variants to `ModelSupportedEndpoint` enum |
| `crates/copilot_chat/src/copilot_chat.rs` | Removed `.dedup_by()` call
that incorrectly filtered models |
| `crates/copilot_chat/src/copilot_chat.rs` | Removed unused
`itertools::Itertools` import |
| `crates/copilot_chat/src/copilot_chat.rs` | Added 8 new unit tests |
## Test Coverage
Added 8 new unit tests to ensure the fix works and prevent regression:
| Test | Purpose |
|------|---------|
| `test_models_with_pending_policy_deserialize` | Verifies models with
non-"enabled" policy states deserialize correctly (they're filtered
later) |
| `test_multiple_anthropic_models_preserved` | Verifies multiple Claude
models are not incorrectly deduplicated |
| `test_models_with_same_family_both_preserved` | Verifies models
sharing the same family (e.g., thinking variants) are both preserved |
| `test_mixed_vendor_models_all_preserved` | Verifies models from
different vendors (OpenAI, Anthropic, Google) are all preserved |
| `test_model_with_messages_endpoint_deserializes` | **Critical test**:
Verifies `/v1/messages` endpoint deserializes correctly |
| `test_model_with_unknown_endpoint_deserializes` | Verifies unknown
future endpoints deserialize to `Unknown` variant |
| `test_model_with_multiple_endpoints` | Verifies models with multiple
endpoints deserialize correctly |
| `test_supports_response_method` | Verifies the `supports_response()`
method logic for endpoint routing |
### Test Results
```
running 10 tests
test tests::test_model_with_messages_endpoint_deserializes ... ok
test tests::test_model_with_multiple_endpoints ... ok
test tests::test_model_with_unknown_endpoint_deserializes ... ok
test tests::test_models_with_pending_policy_deserialize ... ok
test tests::test_models_with_same_family_both_preserved ... ok
test tests::test_mixed_vendor_models_all_preserved ... ok
test tests::test_multiple_anthropic_models_preserved ... ok
test tests::test_resilient_model_schema_deserialize ... ok
test tests::test_supports_response_method ... ok
test tests::test_unknown_vendor_resilience ... ok
test result: ok. 10 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out
```
## How to Test Manually
1. Sign in to GitHub Copilot in Zed
2. Open the model picker (Agent panel → model selector dropdown)
3. Verify that Anthropic Claude models appear in the list:
- Claude Sonnet 4
- Claude Opus 4
- Other Claude variants (if enabled in your GitHub Copilot settings)
## Checklist
- [x] Code compiles without errors
- [x] `./script/clippy --package copilot_chat` passes with no warnings
- [x] All unit tests pass
- [x] Change is focused on a single bug fix
- [x] No unrelated refactoring or feature additions
<img width="320" height="400" alt="Screenshot 2026-01-24 at 11 57 21 PM"
src="https://github.com/user-attachments/assets/d5e17e1b-da80-4f4d-a218-d50d35114a21"
/>
Release Notes:
- Fixed Anthropic models not appearing in the Copilot Chat model picker
---------
Co-authored-by: Piotr Osiewicz <24362066+osiewicz@users.noreply.github.com>
Closes#38168Closes#42414
We were already using `ExitStatusExt::from_raw()` to construct an
`ExitStatus`, but we were passing in the exit code from
alacritty_terminal's `ChildExit` event. This worked fine on Windows
where `from_raw()` expects an exit code, but on Unix, `from_raw()`
([read
more](https://doc.rust-lang.org/std/os/unix/process/trait.ExitStatusExt.html#tymethod.from_raw))
expects a raw wait status from `waitpid()` and not an exit code.
When a child process was killed by a signal (e.g., SIGSEGV),
`ExitStatus::code()` returns `None` since only normal exits have an exit
code. This caused the terminal to hang because we weren't properly
detecting the exit.
One fix would have been to remove the dependency on `ExitStatus`
entirely, but using the raw wait status gives us more information, we
can now detect exit codes, signal terminations, and more.
The actual fix was upstream in `alacritty_terminal` to send the raw wait
status instead of just the exit code.
Currently using forked patch
https://github.com/zed-industries/alacritty/tree/v0.16-child-exit-patch
which is based on v0.25.1.
Upstream PR: https://github.com/alacritty/alacritty/pull/8825
Release Notes:
- Fixed terminal hanging when a child process is killed by a signal
(e.g., SIGSEGV from null pointer dereference).
- **title_bar: Extract platform_title_bar from title_bar**
- **file_finder no longer depends on search and agent_servers no longer
depend on language_models**
Release Notes:
- N/A
---------
Co-authored-by: Zed Zippy <234243425+zed-zippy[bot]@users.noreply.github.com>
Users had trouble signing in due to us relying on the Copilot::global
being set, which was never the case. We've decided to use a dedicated
LSP instance just for handling auth of Copilot Chat and other goodies.
That instance is subscribed to by local Copilot instances for projects.
When the Auth instance changes it's state, local instances are prompted
to re-check their own sign in status.
Closes#47352
Co-authored-by: dino <dinojoaocosta@gmail.com>
Release Notes:
- Fixed authentication issues with Copilot.
---------
Co-authored-by: dino <dinojoaocosta@gmail.com>
Co-authored-by: Zed Zippy <234243425+zed-zippy[bot]@users.noreply.github.com>
We might interface with the LSP using URL heres across remotes that have
differing path styles which then breaks every now and then when we have
windows to unix connections. This should helps us fix these occurences
more correctly
Release Notes:
- N/A *or* Added/Fixed/Improved ...
Closes#26866
### Summary
Adds a git graph to Zed, accessible via the `git_graph::Open` action if
a project has an active repository. There's still more to do, but this
is a solid foundation to expand upon. The code structure is in line with
Zed's codebase and shouldn't require architectural changes to add
missing features.
The git graph can be opened via the command palette (`git graph: open`)
or by binding a key to `git_graph::Open`. It's available when the
project has an active git repository.
### Architecture
Similar to the Debugger, the git graph is split between a data layer and
a view/UI layer. When the view layer is rendering, it queries the data
layer for its active state. This setup allows the data layer to lazily
request graph data (only when needed for rendering), abstracts collab
from the view layer, allows most of the data loading to happen on a
background thread, and makes caching easy to implement.
#### Graph Loading
The graph data is loaded in two phases:
1. `Repository::graph_data()` streams commit structure (SHA, parents,
refs) in chunks of 1000 via `git log`
2. `CommitDataReader` lazily fetches full commit details (author,
timestamp, subject) on-demand using a persistent `git cat-file --batch`
process
This two-phase approach makes the initial loading of the graph as fast
as possible, because `git log` takes significantly longer when all the
needed graph data is queried through it. Zed then lazily loads commits
in the user's viewport through `cat-file --batch`. This makes scrolling
to any place in the graph extremely snappy and benefits the
collaborative architecture by only fetching data needed to render the
graph. It also allows Zed to share commit data between different graph
visualizations (e.g., date order vs. topological order).
#### Performance
Tested on both the Zed and LLVM repositories with good performance in
both cases. The two-phase loading approach and lazy fetching keep the UI
responsive even with large commit histories.
#### Testing
I added property testing that builds randomized commit graphs and
verifies that the graph is constructed correctly. This also works as an
integration test and will be expanded in the future to test collab graph
visualization, graph filtering, commit actions, etc.
### New Crate
- `git_graph` (GPL-licensed) — contains UI and graph computation logic
### Not Yet Implemented
- Remote repository support (collab)
- Filtering by branch
- Commit actions (checkout, cherry-pick, etc.)
- Search
- Open commit view for selected commit
- Resizable columns
- Column filtering
#### Reference
<img width="1624" height="976" alt="Screenshot 2025-01-22 at 8 15 39 PM"
src="https://github.com/user-attachments/assets/0f10924a-3964-462f-b320-42d84d02f7bf"
/>
Special thanks to [Alberto Slavica](https://github.com/pyundev) for
submitting #44405, which was a good base to work off of.
Release Notes:
- git: Add initial version of git graph
---------
Co-authored-by: pyundev <pyundev@users.noreply.github.com>
Co-authored-by: Cole Miller <cole@zed.dev>
Co-authored-by: Zed Zippy <234243425+zed-zippy[bot]@users.noreply.github.com>
This PR adds a thinking toggle for controlling whether to use thinking
for a model in the Zed provider:
<img width="645" height="142" alt="Screenshot 2026-01-22 at 12 34 01 PM"
src="https://github.com/user-attachments/assets/9aa543fe-e708-4840-8b38-1a6fbcb78388"
/>
Previously we would create separate "Thinking" variants of the models
that supported thinking in the model selector.
This only applies to Anthropic models in the Zed provider, currently.
This is gated behind the `cloud-thinking-toggle` feature flag.
Release Notes:
- N/A
---------
Co-authored-by: Neel <neel@zed.dev>
- **settings: Do not leak type serialization into downstream crates**
- **worktree: Separate integration tests into separate target**
Release Notes:
- N/A
This PR implements a UI for the side-by-side diff, using blocks to align
the two sides and adding a coherent `SplitEditorElement`.
Release Notes:
- N/A
---------
Co-authored-by: cameron <cameron.studdstreet@gmail.com>
Co-authored-by: Anthony Eid <hello@anthonyeid.me>
Co-authored-by: Zed Zippy <234243425+zed-zippy[bot]@users.noreply.github.com>
## Summary
Implements user-defined slash commands for the agent panel. Users can
create markdown files in `~/.config/zed/commands/` (user-wide) or
`.zed/commands/` (project-specific) that expand into templated text when
invoked via `/command_name` in the chat interface.
## Features
- **File-based commands**: Create `.md` files that become slash commands
- **Template expansion**: Use `$1`, `$2`, etc. for positional arguments,
or `$ARGUMENTS` for all args
- **Namespacing**: Subdirectories create namespaced commands (e.g.,
`frontend/component.md` → `/frontend:component`)
- **Project & user scopes**: Project commands in `.zed/commands/`, user
commands in config dir
- **Claude compatibility**: Can symlink `~/.claude/commands/` for
compatibility
- **Caching**: Commands are cached and watched for file changes
- **Error handling**: Graceful degradation with error display in UI
## Feature Flag
Behind `user-slash-commands` feature flag.
## Testing
52 unit/integration tests covering parsing, validation, expansion, file
loading, symlinks, and error handling.
---
Release Notes:
- N/A (behind feature flag)
---------
Co-authored-by: Zed Zippy <234243425+zed-zippy[bot]@users.noreply.github.com>
This PR removes the `open-ai-responses-api` feature flag and makes it so
all OpenAI requests to the Zed provider use the Responses API.
We've been running this in Nightly/Preview for a week now without any
issues.
Closes CLO-104.
Release Notes:
- N/A
Closes#47098
The root cause of this issue is related to how `Poetry` (and the
upstream `pet-poetry` library) handles path hashing. While perhaps it's
an upstream behavior, we can easily fix it on the Zed side.
The related code are
https://github.com/zed-industries/zed/blob/23c7d60f1e9a1f9e4ce1502aad3bf90c8093833e/crates/languages/src/python.rs#L1181-L1211
In my debugging, I found that `worktree_root` takes the form
`/home/user/project`, but `config.workspace_directories` often ends up
as `/home/user/project/`(with a trailing slash). Normally this wouldn't
be an issue, but `Poetry` generates environment names based on the hash
of the absolute path. Since the hashes for `/home/user/project` and
`/home/user/project/` are different, `pet-poetry` fails to find the
environment.
The fix is straightforward: we just need to ensure the trailing `/` is
removed so the hashes match.
Release Notes:
- Fixed poetry environment not discovered on linux
Closes#46502
The issue here was that we were not looking into the sub page stack when
looking headers up, resulting in an out of bounds index. This PR fixes
this.
Due to me also fixing another small bug in the UI (and adding proper
support for the breadcrumbs), I had to move quite some stuff around here
to get this to work. Namely, I made the `sub_page_stack` a field on the
`SettingsWindow` and now only store the `active_language` in a global to
ensure that we store scroll positions properly for all sub pages. For
that to work with the edit prediction provider page, I had to remove the
struct there and could just move that into a method, which was a nice
side effect there I suppose.
Release Notes:
- Fixed a crash that could occur when trying to edit edit prediction
providers in the settings UI.
While at it, annotate more functions that are potentially related to
language parsing in buffers.
Also, on macOS, in order to actually have callstack frames properly
recorded by Tracy, you need to manually run `dsymutil` on the binary.
Release Notes:
- N/A
Adds the ability to create a dev container definition from scratch.
Additionally, separates devcontainer logic out into its own crate, since
it was getting sufficiently complex to separate from the
`recent_projects` crate.
A screen recording of the modal experience:
https://github.com/user-attachments/assets/f6cf95e1-eb7b-4ca3-86c7-c1cbc26ca557
Release Notes:
- Added modal to initialize a dev container definition in the project
with `projects: initialize dev container`
- Added podman support for dev container actions with the `use_podman`
setting
- Improved devcontainer error handling
---------
Co-authored-by: Sam Coward <idoru42@gmail.com>
Rustfmt surrendered here because that method grew a bit too large. This
splits this up into smaller methods in an effort to combat this and make
the code more maintainable.
Release Notes:
- N/A
Closes #ISSUE
Moves the settings content definitions into their own crate, so that
they are compiled+cached separately from settings, primarily to avoid
recompiles due to changes in gpui. In that vain many gpui types such as
font weight/features, and `SharedString` were replaced in the content
crate, either with `*Content` types for font/modifier things, or
`String`/`Arc<str>` for `SharedString`. To make the conversions easy a
new trait method in the settings crate named `IntoGpui::into_gpui`
allows for `into()` like conversions to the gpui types in
`from_settings` impls.
Release Notes:
- N/A
---------
Co-authored-by: Piotr Osiewicz <24362066+osiewicz@users.noreply.github.com>
Static schemas (tasks, snippets, jsonc, keymap, action/*, tsconfig,
package_json, inspector_style) are now computed once on first access
using LazyLock and returned immediately via Task::ready().
These schemas never change at runtime because:
- tasks, snippets, jsonc, inspector_style are derived from static Rust
types
- tsconfig, package_json are bundled JSON files (include_str!)
- keymap and action/* depend only on registered actions, which are
collected at compile/link time via the inventory crate (extensions
cannot add actions)
This eliminates foreground thread blocking for these schema requests.
**New functions added to KeymapFile:**
- `generate_json_schema_from_inventory()`: generates keymap schema
without App context
- `get_action_schema_by_name()`: looks up single action schema without
App context
Release Notes:
- N/A
---------
Co-authored-by: Ben Kunkle <ben@zed.dev>
Closes#40322
When users run multi-line terminal commands, the confirmation dialog was
only showing the first line with '- N more lines' truncated. This meant
dangerous commands like 'rm -rf' or 'sudo' operations could be hidden
from view, creating a security risk.
Now terminal commands always display in full, with scroll support for
very long commands (20+ lines) to keep the confirmation buttons visible.
Other tools continue using truncation for better UI efficiency since
they don't pose the same security concerns.
Release Notes:
- Agent: Fixed terminal command truncation in the agent panel to better
expose long commands (e.g., potentially dangerous multi-line commands
are now fully visible before execution)
---------
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
Co-authored-by: dino <dinojoaocosta@gmail.com>
- **copilot: Fix double lease panic when signing out**
- **Extract copilot_chat into a separate crate**
- **Do not use re-exports from copilot**
- **Use new SignIn API**
- **Extract copilot_ui out of copilot**
Closes#7501
Release Notes:
- Fixed Copilot providing suggestions from different Zed windows.
- Copilot edit predictions now support jumping to unresolved
diagnostics.
Feature flagged for now as we test this out with actual agents to see if
we need to provide any more feedback to the RFD before committing to the
current setup.
Release Notes:
- N/A
This makes sure that all of the work we do for caching/refreshing
session history is reused everywhere we need to access the list.
Very important for external agents so we aren't rerequesting history on
every search or recent list update. Puts the reloading of history in one
place
Release Notes:
- N/A
This one does `fim_prefix`, `fim_middle`, and `fim_suffix` in that
order, in the prompt, instead of putting the current middle last.
Release Notes:
- N/A
---------
Co-authored-by: Agus Zubiaga <agus@zed.dev>
Co-authored-by: Ben Kunkle <ben@zed.dev>
This PR introduces a project dropdown when working with multiple
folders/projects in one workspace. Here are some interaction details
that I hope improves the UX of working on this scenario significantly:
- The dropdown shows the currently "active" project, which is determined
by:
- Either the file you're currently editing
- Or the file you have just recently switched to
- Some example cases:
- If you are focused on file from project A but switch to project B in
the titlebar, nothing happens. However, as soon as you type on the file
from project A, the title bar will update and your active project will
return to being project A.
- If you're focused on file from project A and change tabs to a file
from project B, the title bar will update, showing project B as the
active one.
- The content you'll see in the branch picker will correspond to the
currently active project
- It's still possible to reach the "Recent Projects" picker through the
project dropdown
- It's possible to do all interactions (trigger dropdown, select active
project, and remove project from workspace) with the keyboard
https://github.com/user-attachments/assets/e2346757-74df-47c5-bf4d-6354623b6f47
Note that this entire UX is valid only for a multiple folder workspace
scenario; nothing changes for the single project case.
Release Notes:
- Workspace: Improved the UX of working with multiple projects in the
same workspace through introducing a project dropdown that more clearly
shows the currently active project as well as allowing you to change it.
This fixes an issue that we noticed in particular with Mercury edit
predictions.
* [x] fix storage to not go stale
* [x] exclude excerpts that intersect the cursor excerpt
* [x] see if string representation of excerpts can be cached, to avoid
rebuilding it on every prediction
Release Notes:
- N/A
---------
Co-authored-by: Ben Kunkle <ben@zed.dev>
This should move almost all access of the threadstore behind the trait,
which unlocks external agents supplying the list
Release Notes:
- N/A
---------
Co-authored-by: Zed Zippy <234243425+zed-zippy[bot]@users.noreply.github.com>
The scheduler integration (#44810) removed the special handling for
realtime audio tasks that spawn them on a dedicated thread. This caused
a panic in the Mac dispatcher when RealtimeAudio priority was passed to
the regular dispatch path.
This restores the original behavior: RealtimeAudio tasks are spawned on
a dedicated thread via dispatcher.spawn_realtime(), using a bounded
channel to send runnables to it.
Release Notes:
- N/A
Release Notes:
- N/A
TL;DR: There was a closed issue 7 hours ago that changed ashpd version,
it needed another bump.
A dependency of gpui called ashpd has a fix in v0.12.1, GPUI has the
0.12 version. After forking zed and patching the gpui source to be mine
with 0.12.1 it was able to use GPUI as a dependency and compile the
project. The error I was getting and the recomendation to do this PR can
be seen in this [closed issue from
ashpd](https://github.com/bilelmoussaoui/ashpd/issues/325).
After changing the dependecy I hit cargo check:
```bash
Finished `dev` profile [unoptimized + debuginfo] target(s) in 4m 24s
```
And cargo run:
```bash
Compiling zed v0.220.0 (/home/user/git/xaviduds/zed/crates/zed)
Finished `dev` profile [unoptimized + debuginfo] target(s) in 9m 44s
Running `target/debug/zed`
```
Zed compiled and opened:
<img width="1920" height="1042" alt="image"
src="https://github.com/user-attachments/assets/8bc524d1-bc4a-43f8-9da4-b02faecca30c"
/>
Release Notes:
- N/A
Description:
E.g. Instead of saying:
2026-01-07T09:08:42-06:00 ERROR [crates/snippet_provider/src/lib.rs:46]
failed to parse snippet
Caused by:
expected an integer
Now the logs give more valuable information like:
2026-01-07T10:47:39-06:00 ERROR [crates/snippet_provider/src/lib.rs:48]
invalid snippet in /Users/oscarvarto/Library/Application
Support/Zed/extensions/installed/django-snippets/./snippets/python.json
(re_path)
