## Summary
Routes chatd model calls backed by concrete AI Provider rows through the
in-process aibridge transport by default, with deployment options to use
direct provider routing when AI Gateway is disabled or chat AI Gateway
routing is disabled.
- Splits model routing into common, direct provider, and AI Gateway
paths behind a single deployment-mode entry point.
- Builds chatd models through explicit request, route, and options data.
Active API key attribution is passed explicitly instead of being hidden
inside generic model construction.
- For AI Gateway BYOK routes, resolves the user's provider key in chatd,
forwards it through provider-specific auth headers, and sets
`X-Coder-AI-Governance-Token` to the `delegated` marker so aibridge
preserves those headers while still stripping Coder-specific metadata.
- Keeps central provider credentials and deployment fallback credentials
out of forwarded provider auth headers, so AI Gateway central policy
remains authoritative.
- Redacts delegated provider auth from default string formatting to
avoid accidental plaintext logging of user BYOK credentials.
- Covers selected chat models, advisor overrides, title and quickgen
paths, subagent overrides, computer use model selection, and an
integration-style chat turn through the aibridge transport path.
- Persists initiating API key IDs on chat and queued user messages,
including subagent child messages, and fails closed for AI
Gateway-routed model builds without an active key.
- Removes unused `api_key_id` indexes while keeping the persistence
columns and foreign keys.
- Keeps the deployment option available through config and env parsing,
but hides it from CLI help and generated docs.
- Stabilizes the subagent poll fallback test so background CreateChat
processing cannot win the state transition under slower CI environments.
## Tests
- `go test ./coderd/x/chatd -run
'TestAIGatewayProviderAuthForUser|TestAIGatewayProviderAuthRedactsFormatting|TestResolveModelRouteForConfigAIGatewayProviderAuth|TestAIGatewayModelForwardsProviderAuth|TestProcessChat_AIGatewayRoutingUsesDelegatedAPIKey|TestAwaitSubagentCompletion'
-count=1`
- `go test ./coderd/aibridged -run
'TestServeHTTP_DelegatedAPIKey|TestServeHTTP_StripCoderToken' -count=1`
- `git diff --check HEAD~1..HEAD`
- `make lint`
> Mux working on behalf of Mike.
Adds options matching new AI Gateway naming.
New options are added as alias for old options. Old options are still
working.
Old options have deprecated message.
No conflict detection was added.
Updated documentation so it mentions only new options. Added note about
old options still working.
> Various AI tools where used to create this PR
Since Go 1.22, the loop variable capture issue is resolved. Variables
declared by for loops are now per-iteration rather than per-loop, making
the 'v := v' pattern unnecessary.
Closes [#1227](https://github.com/coder/internal/issues/1227)
Added support for license addons, starting with AI Governance, to enable
dynamic feature grouping without requiring license reissuance.
### What changed?
- Introduced a new `Addon` type to represent groupings of features that
can be added to licenses
- Created the first addon `AddonAIGovernance` which includes AI Bridge
and Boundary features
- Added validation for addon dependencies to ensure required features
are present
- Added new features: `FeatureBoundary` and
`FeatureAIGovernanceUserLimit`
- Updated license entitlement logic to handle addons and their features
- Added helper methods to check if features belong to addons
- Updated tests to verify addon functionality
### Why make this change?
This change introduces a more flexible licensing model that allows
features to be grouped into addons that can be added to licenses without
requiring reissuance when new features are added to an addon. This is
particularly useful for specialized feature sets like AI Governance,
where related features can be bundled together and sold as a separate
SKU. The addon approach allows for better organization of features and
more granular control over entitlements.
## Summary
AI Bridge is moving out of Premium as a separate add-on (GA in Feb 3).
Closes https://github.com/coder/internal/issues/1226
## Changes
- Excludes `FeatureAIBridge` from `Enterprise()` and
`FeatureSetPremium.Features()`
- Adds soft warning for deployments with AI Bridge enabled but not
entitled
- Warning is displayed to Auditor/Owner roles in UI banner and CLI
headers
## Warning Message
When AI Bridge is enabled (`CODER_AIBRIDGE_ENABLED=true`) but the
license doesn't include the entitlement:
> AI Bridge has reached General Availability and your Coder deployment
is not entitled to run this feature. Contact your account team
(https://coder.com/contact) for information around getting a license
with AI Bridge.
## Behavior
- The feature remains usable in v2.30 (soft warning only)
- Future versions may include hard enforcement
Closes https://github.com/coder/coder/issues/21360
A few considerations/notes:
- I've kept the number of conns to 10 in all other places, except coderd
- which uses the config value
- I opted to also make idle conns configurable; the greater the delta
between max open and max idle, the more connection churn
- Postgres maintains a [_process_ per
connection](https://www.postgresql.org/docs/current/connect-estab.html),
contrary to what the comment said previously
- Operators should be able to tune this, since process churn can
negatively affect OS scheduling
- I've set the value to `"auto"` by default so it's not another knob one
_has to_ twiddle, and sets max idle = max conns / 3
---------
Signed-off-by: Danny Kopping <danny@coder.com>
Fixes all our Go file imports to match the preferred spec that we've _mostly_ been using. For example:
```
import (
"context"
"time"
"github.com/prometheus/client_golang/prometheus"
"golang.org/x/xerrors"
"gopkg.in/natefinch/lumberjack.v2"
"cdr.dev/slog/v3"
"github.com/coder/coder/v2/codersdk/agentsdk"
"github.com/coder/serpent"
)
```
3 groups: standard library, 3rd partly libs, Coder libs.
This PR makes the change across the codebase. The PR in the stack above modifies our formatting to maintain this state of affairs, and is a separate PR so it's possible to review that one in detail.
**Breaking Change:** Existing oauth apps might now use PKCE. If an
unknown IdP type was being used, and it does not support PKCE, it will
break.
To fix, set the PKCE methods on the external auth to `none`
```
export CODER_EXTERNAL_AUTH_1_PKCE_METHODS=none
```
Add `RetentionConfig` with server flags for configuring data retention:
- `--audit-logs-retention`: retention for audit log entries
- `--connection-logs-retention`: retention for connection logs
- `--api-keys-retention`: retention for expired API keys (default 7d)
Updates #20743
> [!CAUTION]
> In whichever release this lands, we've removed the ability to provide
keys via a YAML file (specifically on `openai_key`, `anthropic_key`,
`bedrock_access_key` and finally `bedrock_access_key_secret`). This will
need to be described in the release notes as to not break peoples AI
Bridge integrations upgrading from older versions.
This pull-request ensures that we can see the overview of the settings
of the `AI Bridge` feature within the `/deployment/observability` route.
This set of options only render when the `aibridge` feature flag is
enabled.
### Preview
Solves #15575
Adds OAuth access token revocation when unlinking external auth
provider. Due to revocation not being consistently implemented by
providers this is only best effort attempt. Unsuccessful revocation
won't influence link removal.
Note that enforcement and checking usage will come in a future PR.
This feature is implemented differently than existing features in a few
ways.
It's highly recommended that reviewers read:
- This document which outlines the methods we could've used for license
enforcement:
https://www.notion.so/coderhq/AI-Agent-License-Enforcement-21ed579be59280c088b9c1dc5e364ee8
- Phase 0 of the actual RFC document:
https://www.notion.so/coderhq/Usage-based-Billing-AI-b-210d579be592800eb257de7eecd2d26d
### Multiple features in the license, a single feature in codersdk
Firstly, the feature is represented as a single feature in the codersdk
world, but is represented with multiple features in the license.
E.g. in the license you may have:
{
"features": {
"managed_agent_limit_soft": 100,
"managed_agent_limit_hard": 200
}
}
But the entitlements endpoint will return a single feature:
{
"features": {
"managed_agent_limit": {
"limit": 200,
"soft_limit": 100
}
}
}
This is required because of our rigid parsing that uses a
`map[string]int64` for features in the license. To avoid requiring all
customers to upgrade to use new licenses, the decision was made to just
use two features and merge them into one. Older Coder deployments will
parse this feature (from new licenses) as two separate features, but
it's not a problem because they don't get used anywhere obviously.
The reason we want to differentiate between a "soft" and "hard" limit is
so we can show admins how much of the usage is "included" vs. how much
they can use before they get hard cut-off.
### Usage period features will be compared and trump based on license
issuance time
The second major difference to other features is that "usage period"
features such as `managed_agent_limit` will now be primarily compared by
the `iat` (issued at) claim of the license they come from. This differs
from previous features. The reason this was done was so we could reduce
limits with newer licenses, which the current comparison code does not
allow for.
This effectively means if you have two active licenses:
- `iat`: 2025-07-14, `managed_agent_limit_soft`: 100,
`managed_agent_limit_hard`: 200
- `iat`: 2025-07-15, `managed_agent_limit_soft`: 50,
`managed_agent_limit_hard`: 100
Then the resulting `managed_agent_limit` entitlement will come from the
second license, even though the values are smaller than another valid
license. The existing comparison code would prefer the first license
even though it was issued earlier.
### Usage period features will count usage between the start and end
dates of the license
Existing limit features, like the user limit, just measure the current
usage value of the feature. The active user count is a gauge that goes
up and down, whereas agent usage can only be incremented, so it doesn't
make sense to use a continually incrementing counter forever and ever
for managed agents.
For managed agent limit, we count the usage between `nbf` (not before)
and `exp` (expires at) of the license that the entitlement comes from.
In the example above, we'd use the issued at date and expiry of the
second license as this date range.
This essentially means, when you get a new license, the usage resets to
zero.
The actual usage counting code will be implemented in a follow-up PR.
### Managed agent limit has a default entitlement value
Temporarily (until further notice), we will be providing licenses with
`feature_set` set to `premium` a default limit.
- Soft limit: `800 * user_limit`
- Hard limit: `1000 * user_limit`
"Enterprise" licenses do not get any default limit and are not entitled
to use the feature.
Unlicensed customers (e.g. OSS) will be permitted to use the feature as
much as they want without limits. This will be implemented when the
counting code is implemented in a follow-up PR.
Closes https://github.com/coder/internal/issues/760
Resolves https://github.com/coder/coder/issues/15513
Disables notifications when both `$CODER_NOTIFICATIONS_WEBHOOK_ENDPOINT` and `$CODER_EMAIL_SMARTHOST` are unset.
Breaking change: `$CODER_EMAIL_SMARTHOST` is no longer set by default as `localhost:587`, meaning any deployments that make use of this default value will need to add it back.
---------
Co-authored-by: Danny Kopping <danny@coder.com>
Co-authored-by: Mathias Fredriksson <mafredri@gmail.com>
* fix(coderd): mark provisioner daemon psk as secret
Marks provisioner daemon PSK with the secret annotation.
This ensures it will be scrubbed from API requests to
/api/v2/deployment/config.
* make gen
* fix: strip timezone information from a date in dau response
Timezone information is lost, so do not forward it to the client.
* fix: timezone offset should be flipped
* Make tests deterministic
* Adds an annotation format_duration_ns to all deployment values of type clibase.Duration
* Adds a unit test that complains if you forget to add the above annotation to a clibase.Duration
* Modifies optionValue() to check for the presence of format_duration_ns when displaying an option.
* feat: allow external services to be authable
* Refactor external auth config structure for defaults
* Add support for new config properties
* Change the name of external auth
* Move externalauth -> external-auth
* Run gen
* Fix tests
* Fix MW tests
* Fix git auth redirect
* Fix lint
* Fix name
* Allow any ID
* Fix invalid type test
* Fix e2e tests
* Fix comments
* Fix colors
* Allow accepting any type as string
* Run gen
* Fix href
See also: https://github.com/coder/coder/pull/9522
- Adds commands `server dbcrypt {rotate,decrypt,delete}` to re-encrypt, decrypt, or delete encrypted data, respectively.
- Plumbs through dbcrypt in enterprise/coderd (including unit tests).
- Adds documentation in admin/encryption.md.
This enables dbcrypt by default, but the feature is soft-enforced on supplying external token encryption keys. Without specifying any keys, encryption/decryption is a no-op.
* chore: add /v2 to import module path
go mod requires semantic versioning with versions greater than 1.x
This was a mechanical update by running:
```
go install github.com/marwan-at-work/mod/cmd/mod@latest
mod upgrade
```
Migrate generated files to import /v2
* Fix gen
Michael Suchacz