# Objective
Document how to use Poolside in Zed through the ACP Registry, Poolside
Agent CLI, manual Custom Agent configuration, and Terminal Threads.
## Solution
- Add Poolside paths to the AI by Company guide.
- Document ACP Registry installation and in-thread login.
- Document CLI-assisted and manual Custom Agent configuration.
- Explain the settings-path and PATH requirements.
- Link to the public Poolside Agent CLI repository and Poolside’s Zed
documentation.
## Testing
- Ran `./script/prettier`.
- Ran `git diff --check`.
- Verified Poolside’s live ACP Registry entry.
- Verified the commands, configuration, and authentication behavior
against the Poolside and Zed implementations.
## Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks are not applicable
- [x] The content follows Zed’s documentation and UI conventions
- [x] Code tests are not applicable to this documentation-only change
- [x] Performance impact is not applicable
---
Release Notes:
- N/A
This adds support for running self-hosted Sweep Next Edit edit
prediction models through Zed's OpenAI-compatible provider. The target
use case is a local deployment based on
[sweepai/sweep-next-edit-1.5B](https://huggingface.co/sweepai/sweep-next-edit-1.5B),
using the rewrite-window prompting approach described in [OSS Next
Edit](https://blog.sweep.dev/posts/oss-next-edit). Personal user testing
shows impressive quality from this small edit prediction model.
This is valuable because it allows more flexibility for Zed users to
self-host state of the art next edit prediction models!
This PR addresses the self-hosted workflow discussed in
https://github.com/zed-industries/zed/discussions/50929.
Note that the self-hosted model path needed more than just a new
dropdown value. Zed has to build the rewrite-window prompt shape the
model expects, map the rewritten window back into anchored edits, and
make `prompt_format: "infer"` work with the filename-style model
identifiers that llama-server local server reports.
The main changes are:
- add a `Sweep` prompt format for OpenAI-compatible edit prediction
providers and route it to a dedicated `SweepPrompt` model path
- build Sweep rewrite-window prompts from the active cursor window,
recent change history, and related file excerpts
- convert rewrite responses back into anchored edits and suppress no-op
rewrites
- add fixed cursor-window extraction used by the Sweep rewrite prompt
path
- recognize `sweep-next-edit` model names during `infer`, including
filename-style identifiers such as
`sweepai_sweep-next-edit-1.5B_sweep-next-edit-1.5b.q8_0.v2.gguf`
- reject reserved Sweep prompt tokens before sending malformed requests
to a self-hosted server
- update the docs to describe the actual Sweep rewrite-window prompt
format
Local setup used for manual verification:
```sh
llama-server \
--hf-repo sweepai/sweep-next-edit-1.5B \
--hf-file sweep-next-edit-1.5b.q8_0.v2.gguf \
--ctx-size 8192 \
--host 127.0.0.1 \
--port 8080
```
```json
{
"edit_predictions": {
"provider": "open_ai_compatible_api",
"open_ai_compatible_api": {
"api_url": "http://127.0.0.1:8080/v1/completions",
"model": "sweepai_sweep-next-edit-1.5B_sweep-next-edit-1.5b.q8_0.v2.gguf",
"prompt_format": "infer",
"max_output_tokens": 512
}
}
}
```
The request that produced this PR included screenshots of the
OpenAI-compatible provider configuration and the prompt format dropdown
with `Sweep` selected.
Tests added in this branch:
- `test_sweep_prompt_format_routes_to_sweep_prompt_model`
- `test_fixed_line_window_around_cursor_start_middle_and_end`
-
`test_sweep_prompt_request_prediction_diffs_rewritten_window_into_anchored_edits`
-
`test_sweep_prompt_request_prediction_returns_none_for_identical_rewrite`
-
`test_original_window_for_current_window_uses_latest_pre_edit_snapshot`
-
`test_original_window_for_current_window_returns_none_without_matching_history`
-
`test_recent_change_block_from_event_formats_original_and_updated_sections`
Verification run locally:
- `cargo test -p zed sweep_prompt_format_routes`
- `cargo test -p zed subscribe_uses_stale_provider_config`
- `cargo test -p edit_prediction sweep_prompt`
- `cargo test -p edit_prediction
fixed_line_window_around_cursor_start_middle_and_end`
- `cargo test -p edit_prediction
test_sweep_prompt_request_prediction_diffs_rewritten_window_into_anchored_edits`
- `cargo test -p edit_prediction
test_sweep_prompt_request_prediction_returns_none_for_identical_rewrite`
- `./script/clippy -p zed -p edit_prediction -p settings_content`
Release Notes:
- Added support for self-hosted Sweep Next Edit models in
OpenAI-compatible edit predictions, including the `sweep` prompt format
and `infer` detection for `sweep-next-edit` model names.
---------
Co-authored-by: Ben Kunkle <ben@zed.dev>
Fixes the bug that made us remove the sandbox.
The bug in question was very dumb:
- there is sophisticated machinery for detecting whether a user-granted
writable path is swapped out for a symlink in the timing gap between
approval and sandbox creation
- there was no equivalent machinery to do the same for the (much larger)
gap between a user *persisting an approval* (either for the current
thread or permanently via settings)
- The fix is essentially to store canonical (i.e. absolute and
symlink-free at all depths) paths as the source of truth, but retain the
raw path for display purposes
- On WSL, there is extra care needed becasue of the bidirectional
mounting (i.e. `/mnt/c/...` and `\\wsl.localhost\Ubuntu\...`). In
particular, `/mnt/c/...` paths, since their inodes do not necessarily
pin NTFS file references, weaken the sandbox guarantees, and so we need
some extra UI to call this out and docs etc...
This also does not remove the feature flag, but just toggles it to
"enabled_for_all"
---
Release Notes:
- N/A or Added/Fixed/Improved ...
---------
Co-authored-by: Richard Feldman <oss@rtfeldman.com>
Co-authored-by: Jakub Konka <kubkon@jakubkonka.com>
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
Co-authored-by: zed-zippy[bot] <234243425+zed-zippy[bot]@users.noreply.github.com>
# Objective
Add a new `agent.compaction_model` setting that lets users specify a
separate language model for context compaction (`/compact` and
auto-compaction), independent of the thread's active conversation model.
Compaction is just summarization — there's no reason to pay Opus prices
for it when a cheaper model does the job faster. We've also seen
reasoning models misbehave on this task (empty responses, repetition
loops at high effort), so picking a dedicated non-reasoning model for
compaction is useful.
## Solution
- New `compaction_model: Option<LanguageModelSelection>` field in
`AgentSettingsContent` and `AgentSettings`, mirroring
`thread_summary_model`.
- New `compaction_model: Option<ConfiguredModel>` slot on
`LanguageModelRegistry` with `select_/set_/compaction_model()` trio,
mirroring the existing pattern. The setter deliberately does **not**
emit a registry event in v1; callers read the slot lazily at compaction
time.
- New `Thread::compaction_model(&self, cx: &App)` helper that returns
the configured model or falls back to `self.model()`. Two call sites —
`Thread::compact` and `perform_compaction_if_needed` — now go through
this helper instead of reading `self.model()` directly.
- `build_compaction_telemetry` accepts the compaction model explicitly
so the `model` field reflects the model that actually streamed the
request. `max_tokens` still derives from `thread.model()` (threshold
semantics are unchanged).
- Documentation updated at `docs/src/ai/agent-settings.md` (new
user-facing setting).
**Example:**
```json
{
"agent": {
"default_model": {
"provider": "anthropic",
"model": "claude-opus-4-6"
},
"compaction_model": {
"provider": "anthropic",
"model": "claude-sonnet-4-5"
}
}
}
```
**Resolution chain:**
```
agent.compaction_model (if set & available)
→ thread.model() (always available, current behavior)
```
**Behavior change:**
| Trigger | Before | After |
| --- | --- | --- |
| Manual `/compact` | Uses `thread.model()` | Uses
`agent.compaction_model` if set & available; else `thread.model()` |
| Auto-compaction | Uses `thread.model()` | Same as above |
| `/compact` when thread has no model | `NoModelConfiguredError` |
Succeeds if `compaction_model` resolves |
| `compaction_model` configured but provider missing / model id unknown
| n/a | Falls back to `thread.model()` and logs a one-time warning |
**Explicit non-goals:**
- No `Event::CompactionModelChanged`(no consumer; the `_cx` parameter on
`set_compaction_model` is intentionally accepted for future use).
- No GUI selector (consistent with all other feature-specific models).
- No per-profile override.
- No runtime API-error fallback — only config-time failure (provider not
registered, model id not in `provided_models`) triggers fallback. This
matches every other feature-specific model.
- No change to threshold calculation, auto-compact trigger, or
`COMPACTION_PROMPT`.
- No propagation to subagent threads.
## Testing
3 new unit tests in `crates/agent/src/thread.rs::tests`:
- `test_compaction_uses_configured_compaction_model` — manual `/compact`
routes to the configured model; thread's primary model receives no
request; telemetry reflects the configured model.
- `test_compaction_falls_back_when_compaction_model_unavailable` —
configured-but-unresolvable falls back to `thread.model()`; telemetry
reflects the fallback model.
- `test_auto_compaction_uses_compaction_model` — auto-compaction
triggered by threshold honors the same setting.
All 14 existing compaction tests still pass. Test suites in
`crates/agent_settings`, `crates/language_model`,
`crates/settings_content`, `crates/agent_ui` unchanged. `cargo clippy`
clean on the changed crates.
**How reviewers can test:**
1. Add `agent.compaction_model` to `settings.json` with a cheaper model
than the thread's primary model, run `/compact`, observe the cheaper
model receives the request.
2. Set `agent.compaction_model` to a non-existent provider/model id, run
`/compact`, observe fallback to thread model and a `log::warn!` line.
3. Trigger auto-compaction by reaching the threshold, observe it uses
`compaction_model`.
**Platforms tested:** local Linux (cargo check + cargo test on agent /
agent_settings / language_model / settings_content / agent_ui crates).
## Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments — no `unsafe`
blocks introduced
- [x] The content adheres to Zed's UI standards — N/A: settings-only
change, no UI touched
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable — model
resolution is an O(1) registry lookup, not on any hot path; no new work
in the streaming loop
Release Notes:
- agent: Add support for specifying which model is used for compaction
(`agent.compaction_model`)
---------
Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
Co-authored-by: Bennet Bo Fenner <bennet@zed.dev>
# Objective
- Adds the three GPT-5.6 models (Sol, Terra, and Luna) that OpenAI now
serves through AWS Bedrock's `bedrock-mantle` endpoint, so they can be
selected as built-in models under the Amazon Bedrock provider.
- Follows up on the native Bedrock Mantle support added in #60480.
- Addresses the feature request in
https://github.com/zed-industries/zed/discussions/61003.
## Solution
- Adds `Gpt5_6Sol`, `Gpt5_6Terra`, and `Gpt5_6Luna` variants to
`MantleModel` in `crates/bedrock/src/models.rs`, with per-model
`id`/`request_id`/`display_name` and shared capability arms.
- All three are Responses-API, `bedrock-mantle`-only models, so they
reuse the existing Mantle request/response plumbing, region gating,
bearer-token auth, and model-picker wiring (`MantleModel::iter()`) with
no other changes required.
- Metadata mirrors the AWS Bedrock model cards
([Sol](https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-openai-gpt-56-sol.html),
[Terra](https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-openai-gpt-56-terra.html),
[Luna](https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-openai-gpt-56-luna.html)):
272K context window, image input, tool and thinking support, and
`openai.gpt-5.6-{sol,terra,luna}` request IDs.
- Updates the Amazon Bedrock section of "Use a Gateway" to mention the
GPT-5.6 family.
## Testing
- `cargo test -p bedrock` — adds `test_gpt_5_6_mantle_model_metadata`
and extends `test_builtin_mantle_models_use_responses_protocol`; all
pass.
- `cargo test -p language_models mantle` — the consumer crate compiles
and all Mantle tests pass.
- `./script/clippy -p bedrock` passes with no new warnings; docs pass
`prettier --check`.
## Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments — N/A, no unsafe
code
- [x] The content adheres to Zed's UI standards — N/A, no UI change
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable
---
Release Notes:
- Added GPT-5.6 Sol, Terra, and Luna models to the Amazon Bedrock
provider via the `bedrock-mantle` endpoint.
---------
Co-authored-by: Anant Goel <anant@zed.dev>
## Summary
- Add an ACP-mode guard to the documented OpenCode bell plugin snippet
- Preserve terminal bell notifications for Terminal Threads while
avoiding writes to stdout when OpenCode is used as an ACP External Agent
## Rationale
The OpenCode bell plugin writes BEL to stdout for Terminal Thread
notifications. When the plugin is installed globally and OpenCode runs
in ACP mode, stdout is the JSON-RPC transport. OpenCode sets
`OPENCODE_CLIENT=acp` in this mode, so the guard prevents BEL bytes from
corrupting ACP JSON-RPC messages, including usage updates and permission
flows.
Release Notes:
- N/A
## Summary
- Document pricing and Zed-hosted context limits for Claude Fable 5,
Claude Sonnet 5, and GPT-5.6 Sol, Terra, and Luna.
- Record recent Gemini and xAI model retirements and replacement
guidance.
- Put the Fable safety-retention warning on the hosted-model reference
and align the privacy docs with the current Anthropic Covered Models
terminology and retention period.
## Sources
- zed-industries/cloud origin/main at
f5d109a868a303241e89ea30d6da8da19699ead4
- OpenAI GPT-5.6 model documentation
- Anthropic Covered Models retention policy
- xAI May 15, 2026 model retirement guide
## Testing
- pnpm dlx prettier@3.5.0 . --check
- mdbook build docs
Release Notes:
- N/A
Closes security loopholes and updates docs:
- installs seccomp filter for blocking naughty syscalls
- tightens macos seatbelt profile
- fetch tool responses that redirect are now constrained by allowed
domains list
Also adds a few "Learn More" buttons that link to the new docs.
Also fixes a bug where the agent would try to create a
`~/.config/zed/AGENTS.md` directory
Also adds unicode confusable detection to URL/path privilege escalation
prompts.
---
Release Notes:
- N/A or Added/Fixed/Improved ...
---------
Co-authored-by: zed-zippy[bot] <234243425+zed-zippy[bot]@users.noreply.github.com>
# Objective
- Describe the objective or issue this PR addresses. -> ** Typo and
grammar fixes!**
## Solution
## Testing
## Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content adheres to Zed's UI standards
([UX/UI](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
and
[icon](https://github.com/zed-industries/zed/blob/main/crates/icons/README.md)
guidelines)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable
## Showcase
---
Release Notes:
- N/A
Adds native support for AWS Bedrock's Mantle endpoint
(`bedrock-mantle`), which serves models with no `Converse`/`Invoke`
support on `bedrock-runtime`, such as GPT-5.5, GPT-5.4, and Grok 4.3 but
more importantly **open-weight** models
Closes#60471
## What's changed
- Renamed the existing `Model` enum in the `bedrock` crate to
`ConverseModel`, and added a new `MantleModel` enum for Mantle-only
models. Mantle models reuse the existing OpenAI-compatible Chat
Completions/Responses request and response plumbing
(`into_open_ai`/`into_open_ai_response`,
`OpenAiEventMapper`/`OpenAiResponseEventMapper`) already used by the
native OpenAI and OpenAI-compatible providers, rather than introducing
new marshalling code.
- Added a `BedrockMantleModel` language model that routes requests to
the `bedrock-mantle` endpoint, dispatching to Chat Completions or the
Responses API depending on the model. Mantle models appear in the model
picker alongside Converse models under the same Bedrock provider.
- Added region gating: `bedrock-mantle` is only available in a subset of
AWS Regions, so using a Mantle model outside of them surfaces a clear
error naming the current Region and the supported ones, instead of an
opaque HTTP failure.
- Implemented Bedrock bearer token authentication for Mantle requests: a
configured Bedrock API key is used as-is, and every other auth method
(IAM credentials, named profile, SSO, automatic) derives a short-term
token by locally SigV4-presigning a `CallWithBearerToken` request. This
requires no extra network round trip and no token caching, since
re-signing locally is cheap.
- Added a specific error for the 403 you get when your credentials have
`bedrock:CallWithBearerToken` but not the separate
`bedrock-mantle:CallWithBearerToken` permission Mantle models require,
since this is the most common misconfiguration.
- Added a `mantle_available_models` setting so custom models served
through `bedrock-mantle` can be configured, the same way other providers
support custom models via `available_models`.
- Documented Mantle models and the new setting in the Amazon Bedrock
section of [Use a
Gateway](https://zed.dev/docs/ai/use-a-gateway#amazon-bedrock).
## Testing
- Added unit tests covering: the local SigV4 bearer-token signing
(including a byte-for-byte cross-check against a reference
implementation), Mantle endpoint URL construction, the
Mantle-supported-regions list, thinking-effort normalization, and the
settings-to-model protocol mapping.
- `cargo test -p bedrock -p language_models -p settings_content -p
settings` passes.
- `./script/clippy` passes with no new warnings.
Release Notes:
- Added native support for AWS Bedrock's Mantle endpoint, enabling
GPT-5.5, GPT-5.4, and Grok 4.3 through the Amazon Bedrock provider.
**TL;DR**: model updates + reasoning levels + fixes discovered when
working on https://github.com/zed-industries/zed/pull/60373
# Objective
Since the model auto-discovery PR was
[cancelled](https://github.com/zed-industries/zed/pull/60373#issuecomment-4886521448),
here is a manual model list update! I also copied the stand-alone
bugfixes/enhancements from that PR.
## Solution
A lot of manual work 😅
**OpenCode Zen**:
- added Fable 5 and Sonnet 5
- added models that were previously only available on OpenCode Go: GLM
5.2, Kimi K2.7 Code, and Minimax M3
- added reasoning levels for all models. I started from the data on
[`Models.dev`](https://models.dev) (the `/api.json` raw data), and then
I matched that with what is shown in the OpenCode CLI and what I know to
be true
**OpenCode Go**:
- added reasoning levels for GLM 5.2
**OpenCode in general**:
- added `protocol` validation for the settings, by moving from a random
`String` to an `enum`, for both nicer error messages (random strings or
typos will get an error instead of using `openai_chat` by default) and
to avoid issues like [folks saying non-existent protocols are a
thing](https://github.com/zed-industries/zed/issues/56869#issuecomment-4550154554)
- enabled parallel tool calls by default. As per [OpenCode developer on
Discord](https://discord.com/channels/1391832426048651334/1471233160993050918/1472020924881702912),
_"almost all models worth using support parallel tool calling
natively"_. Manual tests confirmed all OpenCode Go models support this
correctly (and was enabled by default on the OpenCode side for all but 1
model). I initially wanted to skip this from the release notes, but I
added it so folks are aware of it in case any issues are caused by this
being enabled for all models
- allegedly fixed Google thinking since reasoning levels / thinking
effort levels were added for Google models and an auto-checker LLM
highlighted that was not properly configured
- added support for the new-ish `supports_disabling_thinking` so
thinking-only models don't get a no-impact toggle to disable thinking
I have no idea if any of the Free models will disappear in 2 days or
not, so I did not update those 🤷 (as per decision in
https://github.com/zed-industries/zed/issues/56869#issuecomment-4466637648)
## Testing
The Zen and Google changes were not tested as I don't have a Zen
subscription and I stubbornly refuse to get one.
The Free&Go changes were tested by running a "_rename this variable for
me. add a function. delete the function_" test with a few different
models.
## Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content adheres to Zed's UI standards
([UX/UI](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
and
[icon](https://github.com/zed-industries/zed/blob/main/crates/icons/README.md)
guidelines)
- [ ] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable
---
Release Notes:
- Agent: OpenCode settings now validate `protocol` values
- Agent: OpenCode only shows the "Disable thinking" toggle if thinking
can indeed be disabled/enabled
- Agent: OpenCode models now enable parallel tool calls by default
- Agent: Updated OpenCode Zen models (added Fable 5, Sonnet 5, GLM 5.2,
Kimi K2.7 Code, and Minimax M3)
- Agent: Added OpenCode Go GLM 5.2 reasoning effort levels
- Agent: Added reasoning effort levels for all OpenCode Zen models
- Agent: Fixed thinking for OpenCode Zen Google models
## Summary
- describe agent notifications as OS desktop notifications instead of
claiming a fixed screen position
- keep the agent panel docs accurate across platforms and desktop
environments
Fixes#53588
Release Notes:
- N/A
---------
Co-authored-by: Smit Barmase <heysmitbarmase@gmail.com>
Removes git sandbox feature
The reason is essentially:
- write access to a `.git` dir can be trivially escalated to unsandboxed
access
- therefore, it is misleading to offer git access separate from
unsandboxed access
- instead, we encourage the model to use `--no-optional-locks` to avoid
needing write access to `git status`, etc.
Adds sandboxing to fetch tool
---
Release Notes:
- N/A or Added/Fixed/Improved ...
---------
Co-authored-by: zed-zippy[bot] <234243425+zed-zippy[bot]@users.noreply.github.com>
This makes it so that we show open the settings UI instead of the modal
when adding an MCP server when clicking on `Add server`:
<img width="240" height="320" alt="image"
src="https://github.com/user-attachments/assets/bc5e9059-1053-45d5-8fd5-acfb7da48a35"
/>
We still show the modal when installing an extension. Since we're
hopefully replacing MCP extensions with MCP registry support soon, we
can leave it as is for now.
Release Notes:
- N/A
Closes AI-159
Closes AI-434
Closes AI-435
Release Notes:
- Key agent-related settings now live in the settings editor, close to
all other settings available in Zed. This specifically includes the move
of LLM providers, external agents, and MCP servers to the settings
editor.
---------
Co-authored-by: zed-zippy[bot] <234243425+zed-zippy[bot]@users.noreply.github.com>
Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
Hi there, I'm Celina from Hugging Face! Opening this PR to add
[llama.cpp](https://llama.app) as a model provider
# Objective
Today Zed users running llama.cpp have to fall back to the generic
OpenAI-compatible provider, which means no auto-discovery (the router
mode (`llama serve`) discovers models from the cache and loads them on
demand) and manual configuration of every model and its capabilities.
This PR makes `llama.cpp` a first-class provider with the same
auto-discovery experience.
## Solution
- Add a `llama_cpp` client crate with the OpenAI-compatible chat types
(`/v1/chat/completions`, including `reasoning_content`) and the
discovery types (`/v1/models`, `/props`), mirroring the existing
`ollama` crate.
- Add the provider in
`crates/language_models/src/provider/llama_cpp.rs`, modeled on the
Ollama provider (settings, configuration view, event mapping).
- Auto discover served models and their context length and tool/vision
support from `/props`. Set `auto_discover: false` to list models
manually instead.
- An unloaded model can't be inspected without loading it, so it is
listed with optimistic defaults (large context, tools enabled) and is
usable from the first message; its real context length and tool support
are filled in once it loads. These live behind a shared map, so a model
already selected in an open conversation picks them up without being
re-selected.
- Show load progress. The provider subscribes to `/models/sse` and
surfaces each model's load progress (e.g. "Loading weights 42%") in its
display name, reconciling stale labels against `/v1/models`. Builds
without `/models/sse` degrade gracefully - no progress, and no
capability refresh after the initial discovery.
- Add settings (`api_url`, `auto_discover`, `available_models` with
per-model `max_tokens` / `supports_tools` / `supports_images`,
`context_window`, `custom_headers`), the provider icon, a `default.json`
entry, and documentation under "Use a Local Model".
No new dependencies: the crate reuses existing workspace dependencies,
and shared state uses `std::sync::RwLock`.
## Testing
- Unit tests in both new crates cover wire/response parsing, model
discovery for single-model and router shapes, the cold-start optimistic
defaults, the in-place capability refresh once a model loads, and the
`/models/sse` event handling (state changes, load failure, load
progress).
- Built Zed locally and ran it against a local `llama serve` router:
confirmed models are discovered without manual configuration, that the
first message works before a model has finished loading, that load
progress is shown in the model's display name while it loads, and that
the reported context length and tool support refine to the model's real
values once it finishes loading.
- Platforms: tested on macOS (Apple Silicon).
## Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content adheres to Zed's UI standards
([UX/UI](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
and
[icon](https://github.com/zed-industries/zed/blob/main/crates/icons/README.md)
guidelines)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable
## Showcase
The generation speed (tokens/sec) depends on the machine you're running
the model, here it's a Apple M3 Max 64GB running a 4-bit quant of
https://huggingface.co/Qwen/Qwen3.5-35B-A3B. For the load progress
status, make sure to upgrade your llama.cpp version to the latest build.
https://github.com/user-attachments/assets/0254f6ef-abe9-42ed-810b-ef1a5b8fa3bd
---
Release Notes:
- Added llama.cpp as a language model provider
---------
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
Closes AI-367
This PR significantly refactors the queue feature so that we don't rely
on indexes being counted and synced across various state points of the
thread view. Instead, I'm introducing a dedicated message queue module
where we control through a stable queue entry ID all the states of a
given queued message.
Additionally, it's relevant to note that I changed the default behavior
of queued messages on this PR so that they get sent only at the end of
the generation. This makes the out-of-the-box, default behavior of the
feature consistent across the native and external agents, given we never
could pull off the feature with the latter. I personally never liked
that behavior anyway, but I do appreciate how some people do. Because of
that, though, for the Zed agent, a "Steer" toggle button is available
for whenever you'd like to signal that a queued message should be sent
at the turn boundary. This is not a setting, though; decided to go
simpler at this moment. It's something you can choose on a per-message
basis.
Ultimately, I think this implementation makes the overall shape of the
feature much more stable and easier to maintain overtime.
Release Notes:
- Agent: Changed the default behavior of queued messages for the Zed
agent so that they get send at the end of the generation. The ability to
steer them (i.e., send the message at the end of a turn boundary) is
still possible to be toggled.
# Objective
More and more models are supporting the `max` reasoning effort (DeepSeek
V4, GLM 5.2, etc) and that was missing from Zed.
## Solution
Added `ReasoningEffort::Max` level and implemented support for that
across providers. For most providers this is a no-op as they don't
support that (`max` is missing from `supported_reasoning_levels()`) but
for some providers this required tiny changes: for OpenCode there is now
a proper difference between `xhigh` and `max` (this bug was actually
what triggered this whole PR) and for DeepSeek reasoning levels were
migrated from a custom `deepseek::ReasoningEffort`.
## Testing
Tested and confirmed working with a simple _"rename this variable for
me. add a function. delete the function"_ test across a few providers
(OpenCode, GitHub Copilot, and DeepSeek).
## Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content adheres to Zed's UI standards
([UX/UI](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
and
[icon](https://github.com/zed-industries/zed/blob/main/crates/icons/README.md)
guidelines)
- [ ] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable
## Showcase
N/A
---
Release Notes:
- OpenAI-compatible: added support for `max` reasoning levels
- OpenCode: added support for `max` reasoning levels
This removes sandbox prompt and documentation wording that claimed Git
metadata access exposes the inherited SSH agent socket. Git access
remains limited to protected metadata paths, and the Seatbelt wrapper no
longer carries SSH-agent-specific handling or examples.
Release Notes:
- Fixed AI sandboxing documentation to avoid implying Git metadata
access grants SSH agent access.
# Objective
Fix OpenAI-compatible reasoning/thinking support. Previously,
`OpenAiCompatibleLanguageModel` never reported thinking support, so
configured `reasoning_effort` values did not reliably reach the wire and
Responses API reasoning state could miss `include:
["reasoning.encrypted_content"]`.
Fixes#58289
Addresses #59207 for OpenAI-compatible Responses models configured with
`reasoning_effort`.
## Solution
- Treat a non-`none` configured `reasoning_effort` as OpenAI-compatible
thinking support.
- Expose common OpenAI-style effort levels in the agent UI, using the
configured effort as the default.
- Honor selected reasoning effort for OpenAI-compatible chat-completions
requests.
- Send `reasoning_effort: "none"` when thinking is disabled for a
configured reasoning model.
- Preserve native OpenAI behavior by continuing to use native
model-specific supported efforts and `max_completion_tokens`.
- Add an OpenAI-compatible `max_tokens_parameter` capability for
endpoints that expect output limits as `max_tokens`.
- Parse common streamed thinking fields (`reasoning` and
`reasoning_content`).
- Add provider setup UI and docs for configuring OpenAI-compatible
reasoning models.
## Testing
- `cargo fmt -p language_model_core -p agent_ui -p language_models`
- `cargo test -p language_models provider::open_ai_compatible::tests
--lib`
- `cargo test -p language_models provider::open_ai::tests --lib`
- `cargo test -p open_ai completion::tests --lib`
- `cargo test -p agent_ui
agent_configuration::add_llm_provider_modal::tests --lib`
- `git --no-pager diff --check`
## Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content adheres to Zed's UI standards
([UX/UI](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
and
[icon](https://github.com/zed-industries/zed/blob/main/crates/icons/README.md)
guidelines)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable
---
Release Notes:
- Improved OpenAI-compatible provider setup for reasoning models.
---------
Co-authored-by: Anant Goel <anant@zed.dev>
## Summary
- Adds an `agent.terminal_init_command` setting for Terminal Threads.
- Runs the configured command automatically when creating a new Terminal
Thread, while skipping restored terminal threads.
- Documents the setting and exposes it in the AI settings page.
Closes https://github.com/zed-industries/zed/issues/58697
Closes AI-337
Release Notes:
- Added a setting to automatically run a command when opening a new
terminal thread in the agent panel.
---------
Co-authored-by: Anant Goel <anant@zed.dev>
Audited the AI documentation against the implementation
## Changes
- **tool-permissions**: `skill` tool patterns match the absolute
`SKILL.md` path, not the skill name (fixed the table, example pattern,
and added a clarifying note).
- **external-agents**: Gemini CLI receives Zed's Google AI key when no
env key is set; extension-provided agents are deprecated in favor of the
ACP registry.
- **inline-assistant**: alternatives run alongside the inline assistant
model (not just the default model); fixed a broken anchor and
`gpt-4-mini` → `gpt-5-mini` model id typos.
- **mcp**: corrected the extensions menu item name ("Install New
Servers…"), the remote server `url` example, and removed the nonexistent
`thinking` tool from the profile example.
- **agent-panel**: "New From Summary" lives in the New Thread menu;
corrected the `@`-mention type list (no "instruction files"; added Fetch
and Branch Diff).
- **use-api-access**: DeepSeek default `api_url` includes `/v1`.
- **skills**: folder name is a convention not a requirement; long
descriptions warn rather than fail (measured in bytes); hidden skills
are also `@`-mentionable; dropped the unenforced consecutive-hyphen
rule; reconciled the no-remote-registry note with GitHub import.
- **edit-prediction**: `alt-tab` and `alt-l` are bound across platforms;
predictions are throttled rather than fired on every keystroke.
- **getting-started, windows-and-projects, parallel-agents**: Panel
Layout lives in the title-bar user menu, with
`workspace::UseAgenticLayout`/`UseClassicLayout` action references.
Release Notes:
- N/A
---------
Co-authored-by: zed-zippy[bot] <234243425+zed-zippy[bot]@users.noreply.github.com>
Anthropic now requires limited data retention for models it designates
as covered models (Mythos-class models, such as Claude Fable 5): prompts
and outputs are retained for 30 days for trust and safety purposes, on
every platform where these models are offered, including platforms with
zero-data-retention agreements. See [Anthropic's public
policy](https://support.claude.com/en/articles/15425996-data-retention-practices-for-mythos-class-models).
Our AI privacy docs previously described zero data retention as applying
to all Zed-hosted models unconditionally. This PR updates them to
document the exception:
- `docs/src/ai/privacy-and-security.md`: Adds a "Provider Safety
Retention for Designated Models" section covering what is retained, by
whom, for how long, that no-training commitments still apply, and that
bringing your own key does not avoid retention for covered models.
Qualifies the intro, request-path table, and provider commitments table
accordingly.
- `docs/src/ai/ai-improvement.md`: Qualifies the zero-data-retention
statements and links to the new section.
- `docs/src/business/privacy.md`: Qualifies the zero-data-retention
statement and links to the new section.
Release Notes:
- N/A
Closes AI-344
Closes AI-342
Moves agent skills management out of the agent panel and into the
settings UI. The skill creator (previously its own crate) now lives in
`settings_ui` as a settings page, alongside the skills setup page. The
agent panel's ellipsis menu now links to the settings page instead of
hosting skill management directly.
Release Notes:
- Improved agent skills management by moving it into the settings UI
---------
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
## Summary
- Restructures the AI docs around quick routing, agent paths, provider
access, and company-specific entry points.
- Moves account-owned AI billing/model pages into Account & Billing
while preserving legacy AI redirects.
- Adds focused pages for Zed Agent, external agents, terminal threads,
agent profiles, skills, instructions, API access, subscriptions,
gateways, and local models.
- Updates related links across Git, authentication, business, migration,
and settings reference docs.
## Validation
- `pnpm dlx prettier@3.5.0 . --write`
- `pnpm dlx prettier@3.5.0 . --check`
- `git diff --check -- docs/src`
- Changed-doc local markdown link sanity check
- `mdbook build docs` was attempted; it is currently blocked by an
unrelated existing `docs/src/vim.md:542` action schema error for
`vim::HelixJumpToWord`.
## Follow-up
- Create a ticket to align Zed's in-app AI configuration surfaces with
this documentation model, including Agent Settings, Configure buttons,
Tools, Profiles, MCP, Skills/Instructions, and plain settings.
## Suggested .rules additions
N/A
Release Notes:
- N/A
- add Billing Manager to the organization roles docs with a capability
matrix
- update Business and billing docs to explain non-paid-seat billing
access
- update billing docs for self-serve tax ID updates
Release Notes:
- N/A
Add a `custom_headers` setting to each HTTP-based language model
provider
(Anthropic, Bedrock, DeepSeek, Google, LM Studio, Mistral, Ollama,
OpenAI,
OpenAI-compatible, OpenCode, OpenRouter, Vercel AI Gateway, and xAI) so
users
can attach extra headers to every outgoing request. Headers managed by
Zed
(authentication, content-type, etc.) cannot be overridden and are
skipped with
a warning.
Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable
Release Notes:
- Added support for configuring custom HTTP headers on language model
providers via `language_models.<provider>.custom_headers`.
Signed-off-by: Aurabindo Pillai <mail@aurabindo.in>
This updates docs to reflect changes threshold billing behavior for
individual Zed Pro subscriptions.
- Scope threshold billing language explicitly to individual Zed Pro
subscriptions.
- Remove stale language saying each $10 threshold crossing resets the
threshold to $0.
- Document that threshold invoices start at $10, may rise in $10
increments up to $100, and are not automatically lowered during the same
subscription.
- Split spend-limit docs between Zed Pro and Zed Business, with Business
readers linked to organization billing.
Release Notes:
- N/A
Updates the agent server extension documentation to account for the ACP
server extension deprecation.
Release Notes:
- N/A
---------
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
There's currently no way to customize AI-generated commit messages on a
per-commit basis. You can put instructions in AGENTS.md or a project
rules file, but those get fed into every agent interaction, not just
commit generation. The built-in "Commit message" prompt used to be
editable in the Rules Library, but that was replaced by Skills, which
aren't used when generating commit messages.
This PR adds an `agent.commit_message_instructions` setting. Its
contents are added to the commit message prompt in their own section,
alongside any project rules, so you can ask for a specific format
(Conventional Commits, a ticket prefix, etc.) without changing how the
agent behaves elsewhere:
```json
{
"agent": {
"commit_message_instructions": "Use the Conventional Commits format: <type>(<scope>): <description>."
}
}
```
Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable
Addresses #26503
Release Notes:
- Added an `agent.commit_message_instructions` setting to customize
AI-generated git commit messages.
---------
Co-authored-by: Christopher Biscardi <chris@christopherbiscardi.com>
Added because I'd like to get this skill Danilo made without having to
upload a gist
Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable
Release Notes:
- N/A
<img width="627" height="752" alt="Screenshot 2026-05-28 at 1 20 22 PM"
src="https://github.com/user-attachments/assets/0a7825f0-73c5-49e9-b59a-83924a45de98"
/>
Adds Claude Opus 4.8 for BYOK providers, including Anthropic fast-mode
handling and Bedrock/OpenCode model definitions.
Closes AI-336
Release Notes:
- Added Claude Opus 4.8 BYOK support
Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable
## Summary
- Document that Zed-hosted Gemini models do not use Google context
caching
- Clarify that Gemini usage is billed only as input and output tokens,
with no cached-input price
- Link to Google's Vertex AI context caching and zero-data-retention
documentation for background
## Validation
- `script/generate-action-metadata`
- `mdbook build docs`
Release Notes:
- N/A
Update some content regarding the skills migration and its impact on the
Git commit prompt.
Release Notes:
- N/A
---------
Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
Documents the ways to open the Skill Creator in `docs/src/ai/skills.md`:
- From the Agent Panel using the keybinding ({#kb
agent::OpenRulesLibrary})
- From the Agent Panel `...` menu > **Skills**
- From the command palette via {#action agent::OpenSkillCreator}
Release Notes:
- N/A
Fixes the skill installation instructions in `docs/src/ai/skills.md`.
The previous steps used a `git sparse-checkout` approach that was
inaccurate — the cloned repo would land nested inside the skills folder,
and the sparse-checkout path was wrong for repos that store skills in a
subdirectory. Replaced with a simple instruction to copy the skill's
folder into the appropriate location.
Release Notes:
- N/A
Adds documentation for the Skills settings UI introduced alongside the
Skills feature.
## Changes
- **Updated** `docs/src/ai/skills.md`:
- Updated "Create your own" to lead with `/create-skill` and mention the
Skill Creator UI as a secondary option
- Added "Managing Skills" section documenting the Settings Editor Skills
page (view, open, delete skills)
Release Notes:
- N/A
