feat: byok-observability for aibridge by evgeniy-scherbina · Pull Request #239 · coder/aibridge

evgeniy-scherbina · 2026-03-30T15:54:32Z

Summary

Records credential metadata on each AI Bridge interception, enabling admins to see how requests were authenticated and identify which credential was used.

Each provider's CreateInterceptor now captures two new fields on the interceptor:

credential_kind: centralized, personal_api_key, or subscription
credential_hint: masked credential for audit (e.g. sk-a...(13)...efgh)

Changes

Extended Interceptor interface with SetCredential, CredentialKind, and CredentialHint, backed by an embeddable CredentialFields helper
Each provider sets credential metadata in CreateInterceptor:
- OpenAI: centralized (admin key) or personal (Bearer token)
- Anthropic: centralized, personal_api_key (X-Api-Key), or subscription (Bearer token)
- Copilot: always subscription
Improved MaskSecret to embed hidden character count (e.g. sk-a...(13)...efgh instead of sk-a*************efgh)

Relates to coder/coder#23808

intercept/credential.go

provider/anthropic.go

provider/copilot.go

provider/openai.go

utils/mask.go

intercept/credential.go

intercept/interceptor.go

provider/copilot.go

provider/openai.go

ssncferreira · 2026-04-01T10:58:10Z

provider/openai.go

 	//
 	// In BYOK mode the user's credential is in Authorization. Replace
 	// the centralized key with it so it is forwarded upstream.
+	credKind := intercept.CredentialKindCentralized


Isn't setting credHint missing here? Also, we should probably check if cfg.key is empty (chatgpt case), and error if the credential key is empty.

Isn't setting credHint missing here?

Actually - no, it's later used like this:

interceptor.SetCredential(credKind, utils.MaskSecret(cfg.Key))

It's easier for openai case, because authz header is always used.

if cfg.key is empty (chatgpt case), and error if the credential key is empty

Yeah, I remember about this. I can add in this PR.

@ssncferreira

if cfg.key is empty (chatgpt case), and error if the credential key is empty

We already know that claude (maybe other clients) send HEAD requests without authentication, maybe we don't need it?

Also I think it's not specific to chatgpt case.

My suggestion was to use a credSecret := cfg.Key like we do in Anthropic, instead of setting cfg.Key directly.

Regarding the empty key, I think this case can still happen, where we proceed to the interceptors with an empty key. This would likely be a misconfiguration, e.g., someone using the ChatGPT provider (no key configured) without an actual subscription. I assume in this case, the upstream provider would return a 401, but I think aibridge should catch this early and fail fast rather than sending a request it knows will fail. Claude Code already does this: it errors when no key is configured.

This applies to all providers, not just OpenAI, so I'm fine with addressing it separately 🙂

My suggestion was to use a credSecret := cfg.Key like we do in Anthropic, instead of setting cfg.Key directly.

In Anthropic it's a bit different, we have 2 fields in config:

Key

BYOKBearerToken
So I had to use credSecret := cfg.Key. In OpenAI implementation is easier.

Instead of setting cfg.Key

We set cfg.Key in Anthropic as well.

Regarding the empty key, I think this case can still happen, where we proceed to the interceptors with an empty key.

Theoretically it may be possible. My main concern that as we already know, request without authenticate maybe a valid request (e.g. ClaudeCode sending HEAD healthcheck request).
So I wasn't sure if we want to intervene and fail every request without authentication?

I think for now we need to assume that all requests that reach this point have been authenticated by coder (HEAD requests without auth do not get this far).
The case I was mentioning is the case where the user is correctly authenticated via coder, but not sending any authentication to the upstream provider. This is an edge case, and as we talked yesterday, we can address this after 🙂

@ssncferreira yeah, let's address later

The case I was mentioning is the case where the user is correctly authenticated via coder, but not sending any authentication to the upstream provider.

I meant the same. e.g. agent sent Coder-Token, but didn't send LLM token. I didn't want to fail such request, because it maybe valid request (e.g. healthcheck)?

utils/mask.go

bridge.go

ssncferreira · 2026-04-01T11:17:57Z

I think we are missing adding the credential info to the logs, especially for error cases like 401/403/429

intercept/credential.go

intercept/interceptor.go

provider/anthropic.go

utils/mask.go

utils/mask_test.go

Co-authored-by: Susana Ferreira <susana@coder.com>

provider/anthropic_test.go

utils/mask.go

utils/mask_test.go

evgeniy-scherbina · 2026-04-03T20:37:16Z

@ssncferreira

I think we are missing adding the credential info to the logs, especially for error cases like 401/403/429

I extended logger with credential info in newInterceptionProcessor where we call interceptor.ProcessRequest(rw, r).
It should capture all cases you mentioned.

recorder/types.go

ssncferreira · 2026-04-07T12:24:32Z

provider/openai.go

 	//
 	// In BYOK mode the user's credential is in Authorization. Replace
 	// the centralized key with it so it is forwarded upstream.
+	credKind := intercept.CredentialKindCentralized


My suggestion was to use a credSecret := cfg.Key like we do in Anthropic, instead of setting cfg.Key directly.

Regarding the empty key, I think this case can still happen, where we proceed to the interceptors with an empty key. This would likely be a misconfiguration, e.g., someone using the ChatGPT provider (no key configured) without an actual subscription. I assume in this case, the upstream provider would return a 401, but I think aibridge should catch this early and fail fast rather than sending a request it knows will fail. Claude Code already does this: it errors when no key is configured.

This applies to all providers, not just OpenAI, so I'm fine with addressing it separately 🙂

provider/openai.go

dannykopping

LGTM!

dannykopping · 2026-04-08T06:10:41Z

utils/mask_test.go

-		{"long_api_key", "sk-ant-api03-abcdefgh", "sk-a*************efgh"},
-		{"unicode", "hélloworld🌍!", "hé********🌍!"},
-		{"github_token", "ghp_ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefgh", "ghp_******************************efgh"},
+		{"short", "short", "***"},


Can you please add a follow-up to make these more useful as previously discussed? Not sure what happened to my comment.

I added logging of secret_length as we discussed:

aibridge/bridge.go

Lines 249 to 250 in fefb191

slog.F("credential_hint", cred.Hint),

slog.F("credential_length", cred.Length),

On the call consensus was to use short format to avoid wasting DB space.

If you refer to the fact that whole secret is masked, it's by design? Because secret is too short to reveal smth. (shouldn't happen with real providers).

I think this was your original comment: #239 (comment)

But okay, I'll create follow-up PR on that.

intercept/credential.go

ssncferreira

LGTM 🚀

provider/anthropic_test.go

ssncferreira · 2026-04-08T11:58:23Z

provider/openai.go

 	//
 	// In BYOK mode the user's credential is in Authorization. Replace
 	// the centralized key with it so it is forwarded upstream.
+	credKind := intercept.CredentialKindCentralized


I think for now we need to assume that all requests that reach this point have been authenticated by coder (HEAD requests without auth do not get this far).
The case I was mentioning is the case where the user is correctly authenticated via coder, but not sending any authentication to the upstream provider. This is an edge case, and as we talked yesterday, we can address this after 🙂

ssncferreira · 2026-04-08T12:04:27Z

utils/mask_test.go

+		{"medium_15_chars", "thisisquitelong", "th***ng"},
+		{"long_api_key", "sk-ant-api03-abcdefgh", "sk-a***efgh"},
+		{"unicode", "hélloworld🌍!", "hé***🌍!"},
+		{"github_token", "ghp_ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefgh", "ghp_***efgh"},


small nit: I prefer ... instead of ***. The latter is generally associated with censored words, which is not the case.

okay, I'll replace considering Anthropic and OpenAI use ..., and we use it here: https://github.com/coder/aibridge/blob/main/intercept/apidump/headers.go#L25

I remember @dannykopping proposed ***, so I kept ***.

ssncferreira · 2026-04-08T12:06:48Z

provider/copilot.go

 		ExtraHeaders:   extractCopilotHeaders(r),
 	}

+	cred := intercept.NewCredentialInfo(intercept.CredentialKindBYOK, key)


I'm wondering if it makes sense to add the github token here 🤔 I don't think the copilot case fits the reasoning for logging and debugging on the UI.

@ssncferreira
I had very similar proposal, see: #216 (comment)

provider/openai.go

Co-authored-by: Susana Ferreira <susana@coder.com>

evgeniy-scherbina commented Mar 30, 2026

View reviewed changes

intercept/credential.go Outdated Show resolved Hide resolved

evgeniy-scherbina commented Mar 30, 2026

View reviewed changes

provider/anthropic.go Outdated Show resolved Hide resolved

evgeniy-scherbina commented Mar 30, 2026

View reviewed changes

provider/anthropic.go Show resolved Hide resolved

evgeniy-scherbina force-pushed the yevhenii/byok-observability branch from 9bf93dd to c5dd566 Compare March 30, 2026 22:37

evgeniy-scherbina changed the base branch from yevhenii/aibridge-chatgpt-provider to ssncf/feat-validate-provider-names March 30, 2026 22:38

ssncferreira force-pushed the ssncf/feat-validate-provider-names branch 2 times, most recently from 71c69d3 to 30d306a Compare March 31, 2026 10:48

evgeniy-scherbina force-pushed the yevhenii/byok-observability branch from c5dd566 to d6e95ea Compare March 31, 2026 14:00

ssncferreira force-pushed the ssncf/feat-validate-provider-names branch from 7d0c198 to b1c6f3f Compare March 31, 2026 14:36

evgeniy-scherbina force-pushed the yevhenii/byok-observability branch from d6e95ea to b6a01e9 Compare March 31, 2026 16:47

evgeniy-scherbina changed the base branch from ssncf/feat-validate-provider-names to main March 31, 2026 16:47

evgeniy-scherbina force-pushed the yevhenii/byok-observability branch from e1e15df to b6a01e9 Compare March 31, 2026 17:38

evgeniy-scherbina commented Mar 31, 2026

View reviewed changes

provider/anthropic.go Outdated Show resolved Hide resolved

evgeniy-scherbina force-pushed the yevhenii/byok-observability branch from a610090 to d3e5994 Compare March 31, 2026 19:20

evgeniy-scherbina commented Mar 31, 2026

View reviewed changes

provider/copilot.go Outdated Show resolved Hide resolved

evgeniy-scherbina commented Mar 31, 2026

View reviewed changes

provider/openai.go Outdated Show resolved Hide resolved

evgeniy-scherbina commented Mar 31, 2026

View reviewed changes

utils/mask.go Outdated Show resolved Hide resolved

feat: byok-observability for aibridge

ec6ae45

evgeniy-scherbina force-pushed the yevhenii/byok-observability branch from d3e5994 to ec6ae45 Compare March 31, 2026 19:59

evgeniy-scherbina mentioned this pull request Mar 31, 2026

feat: byok-observability for aibridge coder/coder#23808

Open

evgeniy-scherbina marked this pull request as ready for review March 31, 2026 23:36

ssncferreira reviewed Apr 1, 2026

View reviewed changes

ssncferreira requested a review from dannykopping April 1, 2026 11:02

dannykopping reviewed Apr 1, 2026

View reviewed changes

intercept/credential.go Outdated Show resolved Hide resolved

intercept/interceptor.go Outdated Show resolved Hide resolved

provider/anthropic.go Outdated Show resolved Hide resolved

utils/mask.go Outdated Show resolved Hide resolved

utils/mask_test.go Outdated Show resolved Hide resolved

evgeniy-scherbina and others added 5 commits April 2, 2026 19:25

refactor: re-org credential configuring

8d7a6b6

Update intercept/credential.go

be0295d

Co-authored-by: Susana Ferreira <susana@coder.com>

refactor: minor refactor

08c98cb

refactor: make secret is always called

11709db

refactor: add CredentialKind custom type

74503bd

test: enhance tests

a2e538f

dannykopping reviewed Apr 3, 2026

View reviewed changes

provider/anthropic_test.go Outdated Show resolved Hide resolved

utils/mask.go Outdated Show resolved Hide resolved

utils/mask_test.go Outdated Show resolved Hide resolved

evgeniy-scherbina added 5 commits April 3, 2026 14:05

test: enhance tests

c963324

ci: make fmt

3f63615

fix: revert MaskSecret changes

d6594a2

fix: add small comment to enum

43ee727

feat: log credential info on interception errors

9faf167

evgeniy-scherbina mentioned this pull request Apr 3, 2026

aibridge: Implement Bring Your Own Keys (BYOK) coder/internal#1401

Open

11 tasks

ssncferreira reviewed Apr 7, 2026

View reviewed changes

evgeniy-scherbina added 3 commits April 7, 2026 17:32

refactor: update CredentialKind enum

ec24b93

refactor: MaskSecret should be fixed-length

bd91a8f

feat: keep secret length in logs

fefb191

dannykopping approved these changes Apr 8, 2026

View reviewed changes

ssncferreira approved these changes Apr 8, 2026

View reviewed changes

evgeniy-scherbina and others added 3 commits April 8, 2026 09:44

Update provider/anthropic_test.go

7604eba

Co-authored-by: Susana Ferreira <susana@coder.com>

refactor: minor change in MaskSecret

4e51696

refactor: minor change in MaskSecret

f6b2dff

evgeniy-scherbina merged commit f72a795 into main Apr 8, 2026
4 checks passed

	slog.F("credential_hint", cred.Hint),
	slog.F("credential_length", cred.Length),

Conversation

evgeniy-scherbina commented Mar 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

evgeniy-scherbina Apr 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

ssncferreira commented Apr 1, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

evgeniy-scherbina commented Apr 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

dannykopping left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

evgeniy-scherbina Apr 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ssncferreira left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

evgeniy-scherbina commented Mar 30, 2026 •

edited

Loading

evgeniy-scherbina Apr 3, 2026 •

edited

Loading

evgeniy-scherbina commented Apr 3, 2026 •

edited

Loading

evgeniy-scherbina Apr 8, 2026 •

edited

Loading