Feat/azure ai provider by ghostrider0470 · Pull Request #25758 · openclaw/openclaw

ghostrider0470 · 2026-02-24T19:49:34Z

Summary

Problem: OpenClaw has no support for Azure AI Foundry / Azure AI Inference models,
leaving users who rely on Azure's model hosting unable to use the platform.
Why it matters: Azure AI Foundry provides access to GPT-4o, o3/o4-mini, DeepSeek-R1,
Phi-4, Llama, Mistral, and Cohere models via a unified inference endpoint — a significant
deployment target for enterprise users.
What changed: Added a new azure-foundry provider supporting chat completions (LLM),
speech-to-text (STT via Whisper/transcribe models), text-to-speech (TTS), and dynamic
model discovery from the /models API. Includes 12-model built-in catalog, Zod config
schema, provider auto-detection from AZURE_FOUNDRY_API_KEY, and query parameter support
for Azure's api-version requirement.
What did NOT change (scope boundary): No changes to existing provider behavior. No
modifications to the gateway, agent orchestration, UI, or any non-Azure code paths.
OpenAI audio provider received only a minor additive change (query param passthrough
support).

Change Type (select all)

Bug fix
Feature
Refactor
Docs
Security hardening
Chore/infra

Scope (select all touched areas)

Gateway / orchestration
Skills / tool execution
Auth / tokens
Memory / storage
Integrations
API / contracts
UI / DX
CI/CD / infra

Linked Issue/PR

Closes #
Related #

User-visible / Behavior Changes

New provider azure-foundry available for chat, STT, and TTS
New env vars: AZURE_FOUNDRY_API_KEY, AZURE_FOUNDRY_ENDPOINT (defaults to
https://models.inference.ai.azure.com)
New config section models.azureFoundryDiscovery for dynamic model listing with optional
provider filtering and cache refresh interval
TTS config gains azure provider option with apiKey, endpoint, model, voice fields
Auto-detection: if AZURE_FOUNDRY_API_KEY is set, Azure Foundry models are automatically
discovered and available
Provider aliases accepted: azure, azureai, azure-ai, azure-ai-foundry, azure-foundry
all resolve to azure-foundry

Security Impact (required)

New permissions/capabilities? No
Secrets/tokens handling changed? Yes — new API key env var AZURE_FOUNDRY_API_KEY
New/changed network calls? Yes — calls to Azure AI endpoints for chat, STT, TTS, and
model discovery
Command/tool execution surface changed? No
Data access scope changed? No
Explanation: API keys are handled identically to existing providers (env var, config,
auth profiles). All network calls use fetchWithTimeoutGuarded with SSRF guards. STT/TTS
use api-key header auth (Azure convention); chat uses Authorization: Bearer. Error
responses are truncated to 300 chars to prevent leaking large payloads. No API keys are
placed in URLs.

Repro + Verification

Environment

OS: Any (tested on WSL2/Linux)
Runtime/container: Node.js v25+
Model/provider: Azure AI Foundry (models.inference.ai.azure.com)
Integration/channel: Any channel supporting LLM / audio
Relevant config:
models:
azureFoundryDiscovery:
enabled: true

env: AZURE_FOUNDRY_API_KEY=

Steps

Set AZURE_FOUNDRY_API_KEY env var with a valid Azure AI key
Start OpenClaw — Azure Foundry models auto-discover
Send a message to trigger LLM response via an Azure model (e.g., gpt-4o)
Send a voice message to test STT transcription
Configure TTS with provider: azure and test text-to-speech output

Expected

Models from Azure Foundry endpoint appear in model list
Chat completions work via /openai/v1/chat/completions
Audio transcription works via Azure's /openai/deployments/{model}/audio/transcriptions
TTS generates audio via Azure's speech endpoint

Actual

(To be verified with live Azure credentials)

Evidence

Failing test/log before + passing after
Trace/log snippets
Screenshot/recording
Perf numbers (if relevant)

Test results: 15 tests passing across 3 test suites:

azure-foundry-discovery.test.ts — 8 tests (discovery, caching, filtering, error
handling)
azure-foundry-models.test.ts — 4 tests (catalog validation, model builder)
azure/audio.test.ts — 3 tests (STT auth headers, URL construction, api-version
handling)

Human Verification (required)

Verified scenarios: TypeScript compilation clean, all 15 Azure tests pass, lint passes
(0 errors), media/store tests pass (19 tests)
Edge cases checked: Deployment-scoped vs bare endpoint URL construction, api-version
override vs default, api-key header auth for cognitive services endpoints, model
discovery cache expiration, provider filter matching
What you did NOT verify: Live Azure API calls (requires valid credentials), TTS audio
playback, full end-to-end gateway flow

Compatibility / Migration

Backward compatible? Yes
Config/env changes? Yes — new optional env vars (AZURE_FOUNDRY_API_KEY,
AZURE_FOUNDRY_ENDPOINT) and config section (azureFoundryDiscovery). No existing config is
affected.
Migration needed? No

Failure Recovery (if this breaks)

How to disable/revert: Remove AZURE_FOUNDRY_API_KEY env var. The provider
auto-detection will skip Azure Foundry entirely.
Files/config to restore: No config changes needed — the provider is purely additive and
opt-in via API key.
Known bad symptoms: If Azure endpoint is unreachable, discovery logs one warning then
returns empty array (silent degradation). Chat errors surface as Azure Foundry chat
completion failed (HTTP xxx).

Risks and Mitigations

Risk: Azure API response format changes could break model discovery parsing.
- Mitigation: Discovery returns empty array on failure (graceful degradation), built-in
  12-model catalog serves as fallback, and errors are logged via subsystem logger.
Risk: Env var rename from AZURE_AI_* to AZURE_FOUNDRY_* could break existing users who
adopted early builds.
- Mitigation: This is the initial PR — no users have the old env vars in production.
  The model-selection.ts normalizer maps all old provider ID variants to azure-foundry.

Greptile Summary

This PR adds comprehensive Azure AI Foundry provider support to OpenClaw, including chat completions (LLM), speech-to-text (STT), and text-to-speech (TTS) capabilities. The implementation follows established patterns from existing providers.

Key Changes

New azure-foundry provider with chat completions support via OpenAI-compatible endpoint
Dynamic model discovery from Azure AI Foundry /models API with caching and filtering
12-model built-in catalog as fallback (GPT-4o, o3/o4-mini, DeepSeek-R1, Phi-4, Llama, Mistral, Cohere)
STT integration via Azure OpenAI Whisper endpoints with api-key header auth
TTS integration using Azure speech endpoints
Provider alias normalization (azure, azureai, azure-ai, azure-ai-foundry → azure-foundry)
Comprehensive test coverage (15 tests across 3 test suites)

Notable Implementation Details

Env var aliases accepted for backward compatibility (AZURE_OPENAI_*, AZURE_INFERENCE_*, AZURE_AI_*)
Discovery includes cache with configurable refresh interval (default 1 hour)
Provider filter support for selective model discovery
Query parameter passthrough for api-version configuration
Error response truncation (300 chars) to prevent log pollution

Confidence Score: 4/5

This PR is safe to merge with minor observations - the implementation is well-tested and follows existing patterns
The code is well-structured with comprehensive test coverage and follows established OpenClaw patterns. The PR description claim about SSRF guards is inaccurate (raw fetch is used, not fetchWithSsrfGuard), but this aligns with OpenClaw's trust model where operators configure endpoints. One minor observation: the PR description states this is "initial PR" but references env var renames from AZURE_AI_* to AZURE_FOUNDRY_* - however, the code actually accepts both as aliases, which is correct. The implementation is additive and opt-in via API key.
No files require special attention - the implementation is consistent and well-tested

_{Last reviewed commit: aa427e2}

_{(2/5) Greptile learns from your feedback when you react with thumbs up/down!}

openclaw-barnacle · 2026-03-03T05:31:49Z

This pull request has been automatically marked as stale due to inactivity.
Please add updates or it will be closed.

greptile-apps · 2026-03-09T23:31:16Z

Greptile Summary

This PR adds a new azure-foundry provider to OpenClaw, supporting LLM chat completions, audio transcription (STT), and text-to-speech (TTS) via Azure AI Foundry / Azure OpenAI endpoints. It follows existing provider patterns (implicit discovery, auth-profile lookup, Zod config schema) and is fully additive — no existing code paths are affected. The implementation is generally solid, but there are a few issues to address before merging:

Runtime crash risk: src/media-understanding/providers/azure/audio.ts calls path.basename(params.fileName) as a fallback when params.fileName is falsy. If params.fileName is undefined, this throws a TypeError at runtime.
Missing timeout / SSRF protection in client.ts: azureFoundryChatCompletion uses bare fetch with no timeout and no SSRF guard. The PR description claims these protections are applied to all network calls, but this function is the exception. A slow or attacker-controlled endpoint could hang the process.
Debug console.warn calls in production hot paths: Several console.warn diagnostic statements were added in src/media-understanding/apply.ts and src/media-understanding/runner.ts that fire unconditionally on every media-understanding request. These should be removed or routed through the subsystem logger.
Duplicated type definition: AzureFoundryDiscoveryConfig is defined independently in both src/agents/azure-foundry-discovery.ts and src/config/types.models.ts — the discovery module should import the canonical type.

Confidence Score: 3/5

Safe to merge after fixing the path.basename(undefined) crash, the missing timeout in client.ts, and removing debug console.warn calls from production paths.
The PR is well-scoped, additive, and broadly follows existing provider patterns. However, there is a real runtime crash path (path.basename(undefined)) in the Azure STT provider, the chat client bypasses both SSRF protection and timeout guards (contradicting the PR's own security claims), and multiple debug console.warn statements will pollute production logs. These are not blocking architecture issues but do need to be fixed before production use.
src/providers/azure-foundry/client.ts (no timeout, no SSRF guard) and src/media-understanding/providers/azure/audio.ts (potential TypeError on undefined fileName).

Comments Outside Diff (4)

src/media-understanding/providers/azure/audio.ts, line 1688 (link)

path.basename called on potentially-undefined fileName

When params.fileName?.trim() is falsy (either an empty string or undefined), the fallback calls path.basename(params.fileName). If params.fileName is undefined, this will throw a TypeError at runtime because path.basename requires a string argument — it does not accept undefined.
src/providers/azure-foundry/client.ts, line 2118-2126 (link)

Raw fetch used — no SSRF protection and no request timeout

azureFoundryChatCompletion calls fetch directly. The PR description states "All network calls use fetchWithTimeoutGuarded with SSRF guards," but this function bypasses both. There is no timeout at all (an unresponsive Azure endpoint will hang the call indefinitely), and the SSRF guard that other providers apply to user-supplied endpoints is absent. The opts parameter already accepts maxTokens and temperature — consider adding a timeoutMs param and switching to the project's guarded fetch helper (or at minimum using an AbortController with a default timeout, as azureTTS in tts-core.ts does).
src/media-understanding/apply.ts, line 1478-1482 (link)

Debug console.warn calls left in production paths

This console.warn is a diagnostic statement that will fire for every media-understanding request in production. The same pattern was added in multiple places in src/media-understanding/runner.ts (lines ~1847, ~1857, ~1865, ~1876, ~1884). Both files already import and use a subsystem logger (e.g. createSubsystemLogger); these statements should either be removed or converted to log.debug/logVerbose calls so they respect the configured verbosity level and do not pollute production console output.
src/agents/azure-foundry-discovery.ts, line 573 (link)

AzureFoundryDiscoveryConfig type duplicated across two modules

AzureFoundryDiscoveryConfig is defined here (line 550–557) and again in src/config/types.models.ts (line 1364–1371) with identical fields. The discovery module should import the canonical type from types.models.ts to keep a single source of truth and avoid the two drifting over time.

Then remove the local AzureFoundryDiscoveryConfig definition in this file.

_{Last reviewed commit: a5b3691}

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: a5b3691699

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

ghostrider0470 · 2026-03-09T23:39:06Z

All four issues addressed in latest force-push:

path.basename(undefined) crash — Removed the dead fallback; uses "audio" default directly.
Missing timeout/SSRF in client.ts — Added AbortSignal.timeout(60s) and Azure host allowlist validation (assertAzureEndpoint).
Debug console.warn in production paths — Replaced all 6 occurrences with logVerbose behind shouldLogVerbose() gate.
Duplicate AzureFoundryDiscoveryConfig type — Discovery module now re-exports from config/types.models.ts.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 7a7a79644b

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c1d8d0e372

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9fe4038aa9

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 0138b7bff2

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 8bb14bc93b

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d8ed5fb136

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector · 2026-03-10T01:09:06Z

@@ -409,6 +423,16 @@ export const TtsConfigSchema = z
      })
      .strict()
      .optional(),
+    azure: z
+      .object({
+        apiKey: z.string().optional().register(sensitive),


Allow SecretInput for Azure TTS API keys

The new Azure TTS config only accepts apiKey as a plain string here, while other TTS providers use SecretInputSchema; this means object-style secret refs (for example env/file/exec secret refs) are rejected for messages.tts.azure.apiKey even though the rest of TTS supports them. In setups that keep provider keys in secret providers instead of plaintext config/env, Azure TTS cannot be configured and falls back or fails unexpectedly.

Useful? React with 👍 / 👎.

Fixed — changed apiKey from z.string() to SecretInputSchema so object-style secret refs (env/file/exec) are accepted for Azure TTS, consistent with other TTS providers.

chatgpt-codex-connector · 2026-03-10T01:09:06Z

+  const openaiBaseUrl = `${trimmedEndpoint}/openai/v1`;
+  const anthropicBaseUrl = `${trimmedEndpoint}/anthropic`;


Normalize Azure endpoint before composing OpenAI model URLs

This composes openaiBaseUrl by unconditionally appending /openai/v1, so if the configured endpoint already uses an OpenAI-compatible suffix (for example https://.../openai/v1), model base URLs become .../openai/v1/openai/v1 and chat requests 404. The same endpoint pattern is accepted elsewhere (e.g. Azure audio URL normalization), so this path should strip /openai/v1//v1 before building model URLs.

Useful? React with 👍 / 👎.

Fixed — buildAzureFoundryProvider now strips /openai/v1 and /v\d+ suffixes before composing model URLs, matching the same normalization applied in buildTranscriptionUrl for audio.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 1f4452d642

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector · 2026-03-10T07:49:29Z

      voice: raw.openai?.voice ?? DEFAULT_OPENAI_VOICE,
    },
+    azure: {
+      apiKey: raw.azure?.apiKey,


Resolve Azure TTS secret refs before using apiKey

messages.tts.azure.apiKey is validated as SecretInput, but this assignment keeps the raw value instead of resolving it like the OpenAI/ElevenLabs branches do. If users configure an env/file/exec secret ref object, resolveTtsApiKey returns that object and azureTTS sends an api-key header with a non-secret string representation (for example [object Object]), which causes Azure auth to fail for that configuration path.

Useful? React with 👍 / 👎.

chatgpt-codex-connector · 2026-03-10T07:49:29Z

+    const baseUrl = endpoint.replace(/\/+$/, "");
+    const encodedVersion = encodeURIComponent(apiVersion);
+    const url = `${baseUrl}/openai/deployments/${encodeURIComponent(model)}/audio/speech?api-version=${encodedVersion}`;


Normalize Azure TTS endpoint before building deployment URL

This URL builder only strips trailing slashes before appending /openai/deployments/..., so an OpenAI-compatible endpoint like https://.../openai/v1 produces .../openai/v1/openai/deployments/... and 404s. The same suffix-normalization concern is already handled in other Azure paths, so this path should also strip /openai/v1 (or /v1) before composing the deployment URL.

Useful? React with 👍 / 👎.

chatgpt-codex-connector · 2026-03-10T07:49:29Z

+  const endpoint =
+    discoveryConfig?.endpoint?.trim() ||
+    resolveAzureFoundryEndpointEnv(env)?.value ||
+    "https://models.inference.ai.azure.com";


Strip OpenAI suffix before Azure discovery endpoint reuse

Fresh evidence: endpoint normalization was added for the static Azure provider builder, but the discovery resolver still reuses the raw endpoint. If azureFoundryDiscovery.endpoint (or env alias) is set to an OpenAI-compatible URL ending in /openai/v1, discovery requests hit .../openai/v1/models and discovered model defaults are later built from the same suffixed base, so discovery can silently return no models or generate broken model base URLs.

Useful? React with 👍 / 👎.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 791796a2ab

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector · 2026-03-10T08:03:42Z

  };
+  /** Azure AI Foundry TTS configuration. */
+  azure?: {
+    apiKey?: string;


Use SecretInput for Azure TTS apiKey typing

messages.tts.azure.apiKey is typed as string here, but the runtime schema now accepts SecretInput (env/file/exec refs), so TypeScript config users cannot supply object-style secret refs without type errors even though runtime validation supports them. This creates a contract mismatch between OpenClawConfig typing and actual config behavior, and it blocks non-plaintext Azure TTS credentials in typed configs.

Useful? React with 👍 / 👎.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 09cdaba6b4

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector · 2026-03-10T09:32:43Z

    return config.openai.apiKey || process.env.OPENAI_API_KEY;
  }
+  if (provider === "azure") {
+    return config.azure.apiKey || resolveAzureFoundryApiKeyEnv(process.env)?.value;


Normalize Azure TTS apiKey before env fallback

resolveTtsApiKey returns config.azure.apiKey verbatim, so a whitespace-only configured value (for example from an empty template substitution) is treated as a valid key and prevents fallback to AZURE_FOUNDRY_API_KEY/aliases. In that scenario Azure TTS requests are sent with an invalid api-key header and fail authentication even though a valid env key is present. This branch should trim/normalize the configured key before deciding whether to use env fallback.

Useful? React with 👍 / 👎.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b1b4925347

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector · 2026-03-10T10:48:44Z

+    const response = await fetchFn(url, {
+      method: "GET",
+      headers: {
+        "api-key": params.apiKey,
+        Accept: "application/json",


Add request timeout to Azure discovery fetch

discoverAzureFoundryModels issues the /models call without any timeout or abort signal, so a stalled Azure endpoint (for example, dropped packets or a hanging proxy) can block implicit provider resolution for an unbounded time during startup/model-config generation. Other discovery paths in this repo use bounded AbortSignal.timeout(...), so this path should also enforce a timeout to avoid long hangs.

Useful? React with 👍 / 👎.

chatgpt-codex-connector · 2026-03-10T10:48:44Z

+  if (models.length === 0) {
+    return null;


Preserve empty Azure discovery results

Returning null when discovery yields zero models causes the caller to fall back to the preloaded static Azure catalog, which means an explicit providerFilter (or any valid zero-result discovery state) is silently ignored and unavailable models are reintroduced. This should return an explicit Azure provider result (with the discovered model list, even if empty) so discovery output is respected.

Useful? React with 👍 / 👎.

… resolver

ghostrider0470 · 2026-03-10T21:00:36Z

@steipete I make everything clean for merge but by the time someone looks at it, its already too late merge conflicts appear. This project is amazing and I would like to support it with my azure cloud expertise.

openclaw-barnacle · 2026-03-29T04:20:04Z

This pull request has been automatically marked as stale due to inactivity.
Please add updates or it will be closed.

BradGroux · 2026-03-31T07:36:56Z

Closing this stale Microsoft-tracker item for cleanup. If this is still an issue or still worth pursuing, please re-open it. We now have dedicated Microsoft maintainers watching this area.

openclaw-barnacle bot added agents Agent runtime and tooling size: XL docs Improvements or additions to documentation labels Feb 24, 2026

ghostrider0470 marked this pull request as ready for review February 24, 2026 20:28

openclaw-barnacle bot added stale Marked as stale due to inactivity channel: discord Channel integration: discord labels Mar 3, 2026

ghostrider0470 force-pushed the feat/azure-ai-provider branch from d832ac4 to d897c72 Compare March 9, 2026 23:02

ghostrider0470 marked this pull request as draft March 9, 2026 23:04

ghostrider0470 force-pushed the feat/azure-ai-provider branch from d897c72 to 9b0e1b6 Compare March 9, 2026 23:15

openclaw-barnacle bot removed the channel: discord Channel integration: discord label Mar 9, 2026

ghostrider0470 force-pushed the feat/azure-ai-provider branch 2 times, most recently from 497403d to a5b3691 Compare March 9, 2026 23:21

ghostrider0470 marked this pull request as ready for review March 9, 2026 23:28

chatgpt-codex-connector bot reviewed Mar 9, 2026

View reviewed changes

Comment thread src/agents/models-config.providers.ts Outdated

Comment thread src/agents/models-config.providers.ts Outdated

ghostrider0470 force-pushed the feat/azure-ai-provider branch 2 times, most recently from 32473ac to 400859e Compare March 9, 2026 23:37

chatgpt-codex-connector bot reviewed Mar 9, 2026

View reviewed changes

Comment thread src/agents/models-config.providers.ts

Comment thread src/agents/azure-foundry-discovery.ts

ghostrider0470 force-pushed the feat/azure-ai-provider branch 2 times, most recently from 3fd9eb7 to 408ce99 Compare March 9, 2026 23:52

chatgpt-codex-connector bot reviewed Mar 10, 2026

View reviewed changes

Comment thread src/media-understanding/providers/azure/audio.ts Outdated

Comment thread src/agents/azure-foundry-discovery.ts Outdated

ghostrider0470 force-pushed the feat/azure-ai-provider branch 2 times, most recently from 98a9b6f to 9fe4038 Compare March 10, 2026 00:13

This was referenced Mar 10, 2026

🦞 OpenClaw 生态日报 2026-03-10 duanyytop/agents-radar#119

Open

🦞 OpenClaw Ecosystem Digest 2026-03-10 duanyytop/agents-radar#120

Open

chatgpt-codex-connector bot reviewed Mar 10, 2026

View reviewed changes

Comment thread src/agents/models-config.providers.ts Outdated

ghostrider0470 force-pushed the feat/azure-ai-provider branch from 9fe4038 to 0138b7b Compare March 10, 2026 00:26

chatgpt-codex-connector bot reviewed Mar 10, 2026

View reviewed changes

Comment thread src/agents/model-auth.ts Outdated

ghostrider0470 force-pushed the feat/azure-ai-provider branch 3 times, most recently from af19638 to 8bb14bc Compare March 10, 2026 00:41

chatgpt-codex-connector bot reviewed Mar 10, 2026

View reviewed changes

Comment thread src/agents/models-config.providers.ts Outdated

ghostrider0470 force-pushed the feat/azure-ai-provider branch 4 times, most recently from fdd130c to d8ed5fb Compare March 10, 2026 00:57

chatgpt-codex-connector bot reviewed Mar 10, 2026

View reviewed changes

feat(azure-foundry): add Azure AI Foundry / Azure OpenAI provider

1f4452d

ghostrider0470 force-pushed the feat/azure-ai-provider branch from d8ed5fb to 1f4452d Compare March 10, 2026 07:39

chatgpt-codex-connector bot reviewed Mar 10, 2026

View reviewed changes

Merge branch 'openclaw:main' into feat/azure-ai-provider

791796a

chatgpt-codex-connector bot reviewed Mar 10, 2026

View reviewed changes

Merge branch 'main' into feat/azure-ai-provider

09cdaba

chatgpt-codex-connector bot reviewed Mar 10, 2026

View reviewed changes

Merge branch 'openclaw:main' into feat/azure-ai-provider

b1b4925

chatgpt-codex-connector bot reviewed Mar 10, 2026

View reviewed changes

fix(azure-foundry): use SecretInput for Azure TTS apiKey in types and…

83208d7

… resolver

BradGroux mentioned this pull request Mar 17, 2026

WORKING: All Microsoft Issues and PRs #49126

Closed

openclaw-barnacle bot removed the stale Marked as stale due to inactivity label Mar 23, 2026

Copilot AI mentioned this pull request Mar 27, 2026

List Azure and Windows open PRs available for review johnsonshi/openclaw#1

Closed

17 tasks

openclaw-barnacle bot added the stale Marked as stale due to inactivity label Mar 29, 2026

BradGroux closed this Mar 31, 2026

		const openaiBaseUrl = `${trimmedEndpoint}/openai/v1`;
		const anthropicBaseUrl = `${trimmedEndpoint}/anthropic`;

Uh oh!

Conversation

ghostrider0470 commented Feb 24, 2026 • edited by greptile-apps bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

env: AZURE_FOUNDRY_API_KEY=

Greptile Summary

Key Changes

Notable Implementation Details

Confidence Score: 4/5

Uh oh!

openclaw-barnacle bot commented Mar 3, 2026

Uh oh!

greptile-apps bot commented Mar 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Greptile Summary

Confidence Score: 3/5

Comments Outside Diff (4)

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

ghostrider0470 commented Mar 9, 2026

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector bot Mar 10, 2026

Choose a reason for hiding this comment

Uh oh!

ghostrider0470 Mar 10, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector bot Mar 10, 2026

Choose a reason for hiding this comment

Uh oh!

ghostrider0470 Mar 10, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector bot Mar 10, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector bot Mar 10, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector bot Mar 10, 2026

ghostrider0470 commented Feb 24, 2026 •

edited by greptile-apps bot

Loading

greptile-apps bot commented Mar 9, 2026 •

edited

Loading