refactor: remove Amazon Bedrock as first-class provider by cv · Pull Request #1212 · NVIDIA/NemoClaw

cv · 2026-03-31T22:35:30Z

Summary

Reverts #963 and adds a guard test that prevents new CSP-specific provider keys from being added to the installer.

Rationale

Bedrock's OpenAI-compatible endpoint already works through the existing compatible-endpoint option — the same path every other CSP wrapper (Vertex, Azure OpenAI, etc.) should use. Adding a dedicated provider key for each CSP does not scale: each one requires config in onboard.js, a case in inference-config.js, UX menu entries, model curation, validation logic, tests, and docs. The compatible-endpoint and compatible-anthropic-endpoint options already cover these use cases generically.

Changes

bin/lib/onboard.js: Remove BEDROCK_ENDPOINT_URL, bedrock from REMOTE_PROVIDER_CONFIG, REMOTE_MODEL_OPTIONS, validProviders, interactive menu, region check, validation, setupInference, and dashboard label.
bin/lib/inference-config.js: Remove bedrock case from getProviderSelectionConfig().
test/inference-config.test.js: Remove Bedrock-specific assertions; add does not grow beyond the approved provider set guard test that fails if any new provider key (bedrock, vertex, azure, etc.) is introduced.

Type of Change

Code change for a new feature, bug fix, or refactor.

Checklist

Formatters applied
Tests added (guard test)
No secrets committed
npm test passes (725 passed)

Summary by CodeRabbit

Release Notes

Chores
- Removed Amazon Bedrock from the list of available inference providers
- All Bedrock-related configuration, endpoint setup, and onboarding support have been completely eliminated from the system
- Provider selection now enforces an approved provider allowlist to improve system reliability and compatibility
- Validation updated to ensure only approved providers can be configured and used

Bedrock's OpenAI-compatible endpoint works through the existing "compatible-endpoint" option — the same path every other CSP wrapper (Vertex, Azure OpenAI, etc.) should use. Adding a dedicated provider key for each CSP does not scale and creates maintenance overhead for config, tests, UX, and docs that is better avoided. Also adds a guard test that enumerates the approved provider set and fails if a new provider key is introduced, steering contributors toward the compatible-endpoint / compatible-anthropic-endpoint options instead. Reverts: #963 Closes: #1204

coderabbitai · 2026-03-31T22:35:42Z

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 29378cdd-59cd-4bc7-b522-5b8be8829e95

📥 Commits

Reviewing files that changed from the base of the PR and between 75b8bb9 and 8a1a058.

📒 Files selected for processing (3)

bin/lib/inference-config.js
bin/lib/onboard.js
test/inference-config.test.js

📝 Walkthrough

Walkthrough

This pull request removes Amazon Bedrock inference provider support from the codebase. Bedrock configuration handling is removed from the provider selection function, onboarding flows, and model options. Tests specific to Bedrock are deleted, with new validation added to enforce an approved providers list.

Changes

Cohort / File(s)	Summary
Provider Configuration Removal `bin/lib/inference-config.js`	Deleted `case "bedrock"` branch from `getProviderSelectionConfig()`, causing Bedrock provider to return `null` instead of endpoint configuration.
Onboarding Integration Removal `bin/lib/onboard.js`	Removed Bedrock endpoint URL construction, model options from `REMOTE_MODEL_OPTIONS`, `bedrock` from `REMOTE_PROVIDER_CONFIG` and provider allowlists, Bedrock-specific region/validation logic, and provider routing for inference setup.
Test Updates `test/inference-config.test.js`	Removed Bedrock-specific test assertions for config and default model; added new validation test enforcing an approved providers list with rejection of non-approved providers like `bedrock`, `vertex`, and `azure*`.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

refactor: remove Amazon Bedrock as first-class provider #1212 — Parallel PR making identical Bedrock removal changes across the same three files with matching line modifications.

Suggested reviewers

kjw3

Poem

🐰 The Bedrock has left our humble ground,
No Amazon traces to be found,
We've cleaned the config, tests, and more,
The provider registry stands at its door,
Hopping forward, sleek and light! 🏃‍♂️✨

🚥 Pre-merge checks | ✅ 3 | ❌ 2

❌ Failed checks (2 warnings)

Check name	Status	Explanation	Resolution
Linked Issues check	⚠️ Warning	The PR objectives contradict the linked issue `#1204`, which requests adding Bedrock tests. Instead, this PR removes Bedrock support entirely, which is opposite to the linked issue's goal.	Clarify the relationship: this PR reverts `#963` and closes `#1204` as obsolete. Update documentation or linked issues to reflect this is a reversal of prior decisions.
Docstring Coverage	⚠️ Warning	Docstring coverage is 40.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and specifically describes the main change: removing Amazon Bedrock as a first-class provider, which aligns with the primary objective across all modified files.
Out of Scope Changes check	✅ Passed	All changes consistently support the stated objective of removing Bedrock as a first-class provider. No unrelated or out-of-scope changes detected in the modified files.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

Create stacked PR
Commit on current branch

🧪 Generate unit tests (beta)

Create PR with unit tests
Commit unit tests in branch revert/bedrock-provider

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

github-actions · 2026-03-31T22:36:08Z

🚀 Docs preview ready!

https://NVIDIA.github.io/NemoClaw/pr-preview/pr-1212/

## Summary Reverts #963 and adds a guard test that prevents new CSP-specific provider keys from being added to the installer. ## Rationale Bedrock's OpenAI-compatible endpoint already works through the existing `compatible-endpoint` option — the same path every other CSP wrapper (Vertex, Azure OpenAI, etc.) should use. Adding a dedicated provider key for each CSP does not scale: each one requires config in `onboard.js`, a case in `inference-config.js`, UX menu entries, model curation, validation logic, tests, and docs. The `compatible-endpoint` and `compatible-anthropic-endpoint` options already cover these use cases generically. ## Changes - **`bin/lib/onboard.js`**: Remove `BEDROCK_ENDPOINT_URL`, `bedrock` from `REMOTE_PROVIDER_CONFIG`, `REMOTE_MODEL_OPTIONS`, `validProviders`, interactive menu, region check, validation, `setupInference`, and dashboard label. - **`bin/lib/inference-config.js`**: Remove `bedrock` case from `getProviderSelectionConfig()`. - **`test/inference-config.test.js`**: Remove Bedrock-specific assertions; add `does not grow beyond the approved provider set` guard test that fails if any new provider key (bedrock, vertex, azure, etc.) is introduced. ## Related - Reverts #963 - Closes #1204 ## Type of Change - [x] Code change for a new feature, bug fix, or refactor. ## Checklist - [x] Formatters applied - [x] Tests added (guard test) - [x] No secrets committed - [x] `npm test` passes (725 passed)  ## Summary by CodeRabbit * **Breaking Changes** * Removed support for Amazon Bedrock as an inference provider. * **Documentation** * Updated command documentation to recommend `nemoclaw onboard` for new installations. * Marked legacy `nemoclaw setup` command as deprecated. * Updated deployment guides to reference the `nemoclaw onboard` setup wizard. * **Tests** * Updated test suite to validate approved inference providers.  Co-authored-by: Deepak Jain <deepujain@users.noreply.github.com>

VKudlay · 2026-04-02T21:59:01Z

Hey @cv. I feel like the core contribution of the original push was to allow converse api (though I could be wrong). I believe Anthropic over Bedrock is only supported via Converse and InvokeModel API.

(I'm also having a miserable time validating this since AWS docs are super-opaque on this. Please correct me if I'm wrong, but I think that's why bedrock is generally treated separately and may be needed.)

EXECUTING

curl -s -X GET https://bedrock-mantle.us-east-1.api.aws/v1/models \
    -H "Authorization: Bearer $AWS_BEARER_TOKEN_BEDROCK" \
    | jq -r '.data[].id'

qwen.qwen3-next-80b-a3b-instruct
deepseek.v3.2
zai.glm-4.6
minimax.minimax-m2.5
minimax.minimax-m2.1
qwen.qwen3-coder-30b-a3b-instruct
writer.palmyra-vision-7b
openai.gpt-oss-20b
mistral.ministral-3-8b-instruct
qwen.qwen3-vl-235b-a22b-instruct
qwen.qwen3-coder-480b-a35b-instruct
zai.glm-5
mistral.devstral-2-123b
mistral.magistral-small-2509
mistral.voxtral-mini-3b-2507
openai.gpt-oss-120b
nvidia.nemotron-nano-12b-v2
nvidia.nemotron-nano-3-30b
zai.glm-4.7-flash
google.gemma-3-27b-it
openai.gpt-oss-safeguard-120b
minimax.minimax-m2
zai.glm-4.7
mistral.voxtral-small-24b-2507
moonshotai.kimi-k2-thinking
google.gemma-3-12b-it
nvidia.nemotron-super-3-120b
moonshotai.kimi-k2.5
qwen.qwen3-235b-a22b-2507
mistral.ministral-3-3b-instruct
qwen.qwen3-32b
nvidia.nemotron-nano-9b-v2
google.gemma-3-4b-it
qwen.qwen3-coder-next
mistral.ministral-3-14b-instruct
mistral.mistral-large-3-675b-instruct
openai.gpt-oss-safeguard-20b
deepseek.v3.1

VKudlay · 2026-04-02T23:45:38Z

Spent a bit more time on this, since some may not have access. Can at least confirm bedrock Anthropic with chat-completions is non-trivial:

Getting List of Available Models

curl -s -X GET "https://bedrock.us-east-1.amazonaws.com/foundation-models" \
  -H "Authorization: Bearer $AWS_BEARER_TOKEN_BEDROCK" \
  -H "Content-Type: application/json" | jq -r '.modelSummaries[].modelId' | grep -E "anthropic|nvidia|openai"

nvidia.nemotron-nano-12b-v2
anthropic.claude-sonnet-4-20250514-v1:0
anthropic.claude-haiku-4-5-20251001-v1:0
openai.gpt-oss-120b-1:0
nvidia.nemotron-nano-3-30b
anthropic.claude-sonnet-4-6
nvidia.nemotron-super-3-120b
anthropic.claude-opus-4-6-v1
nvidia.nemotron-nano-9b-v2
openai.gpt-oss-20b-1:0
openai.gpt-oss-safeguard-120b
anthropic.claude-sonnet-4-5-20250929-v1:0
anthropic.claude-opus-4-1-20250805-v1:0
anthropic.claude-opus-4-5-20251101-v1:0
openai.gpt-oss-safeguard-20b
anthropic.claude-3-sonnet-20240229-v1:0:28k
anthropic.claude-3-sonnet-20240229-v1:0:200k
anthropic.claude-3-sonnet-20240229-v1:0
anthropic.claude-3-haiku-20240307-v1:0:48k
anthropic.claude-3-haiku-20240307-v1:0:200k
anthropic.claude-3-haiku-20240307-v1:0
anthropic.claude-3-7-sonnet-20250219-v1:0
anthropic.claude-3-5-haiku-20241022-v1:0
anthropic.claude-opus-4-20250514-v1:0

Querying Nemotron Super (OK)

curl -X POST "https://bedrock-runtime.us-east-1.amazonaws.com/openai/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AWS_BEARER_TOKEN_BEDROCK" \
  -d '{
    "model": "nvidia.nemotron-super-3-120b",
    "messages": [{"role": "user", "content": "Hello world" }],
    "max_tokens": 10
  }'

{
    "choices":[{"finish_reason":"length","index":0,"logprobs":null,"message":{
        "content":"Hello! 👋 How can I help you", "refusal":null, "role":"assistant"
    }}],
    "created":1775169709,"id":"chatcmpl-bdd230ad-7290-4650-a837-9e6a710e13e0",
    "model":"nvidia.nemotron-super-3-120b","object":"chat.completion","service_tier":"default",
    "usage":{"completion_tokens":10,"prompt_tokens":18,"total_tokens":28}
}

Querying GPT-OSS (OK)

curl -X POST "https://bedrock-runtime.us-east-1.amazonaws.com/openai/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AWS_BEARER_TOKEN_BEDROCK" \
  -d '{
    "model": "openai.gpt-oss-20b-1:0",
    "messages": [{"role": "user", "content": "Hello world" }],
    "max_tokens": 10
  }'

{
    "choices":[{"finish_reason":"length","index":0,"logprobs":null,"message":{
        "content":"The user says \"Hello world\".","refusal":null,"role":"assistant"
    }}],
    "created":1775169815,"id":"chatcmpl-d69bd361-f8a0-4420-8ac5-63b8732b21f3",
    "model":"openai.gpt-oss-20b-1:0","object":"chat.completion","service_tier":"default",
    "usage":{"completion_tokens":10,"prompt_tokens":67,"total_tokens":77}
}

Querying Claude Opus via Chat/Completions (NOT OK)

curl -X POST "https://bedrock-runtime.us-east-1.amazonaws.com/openai/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AWS_BEARER_TOKEN_BEDROCK" \
  -d '{"model": "anthropic.claude-opus-4-6-v1", "messages": [{"role": "user", "content": "Hello world" }], "max_tokens": 10}'

curl -X POST "https://bedrock-runtime.us-east-1.amazonaws.com/openai/v1/chat/completions" \ns" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AWS_BEARER_TOKEN_BEDROCK" \
  -d '{"model": "global.anthropic.claude-opus-4-6-v1", 
  "messages": [{"role": "user", "content": "Hello world" }], "max_tokens": 10}'

{"error":{
    "message":"The model doesn't exist or doesn't support this API. Retry your request with a different model ID.",
    "type":"not_found_error","param":null,"code":"model_not_found"
}}

Querying General Claude Opus via Converse (NOT OK)

curl -X POST "https://bedrock-runtime.us-east-1.amazonaws.com/model/anthropic.claude-opus-4-6-v1/converse" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AWS_BEARER_TOKEN_BEDROCK" \
  -d '{"messages": [{"role": "user", "content": [{"text": "Hello"}]}]}'

{
  "message":"Invocation of model ID anthropic.claude-opus-4-6-v1 with on-demand throughput isn’t supported.
   Retry your request with the ID or ARN of an inference profile that contains this model."
}

Querying Regional Claude Opus via Converse (OK)

curl -X POST "https://bedrock-runtime.us-east-1.amazonaws.com/model/anthropic.claude-opus-4-6-v1/converse" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AWS_BEARER_TOKEN_BEDROCK" \
  -d '{"messages": [{"role": "user", "content": [{"text": "Hello"}]}]}'

{"metrics":{"latencyMs":1824},"output":{"message":{"content":[{
    "text":"Hello! How are you doing today? Is there anything I can help you with?"
  }],"role":"assistant"}},"stopReason":"end_turn","usage":{
    "cacheReadInputTokenCount":0,"cacheReadInputTokens":0,
    "cacheWriteInputTokenCount":0,"cacheWriteInputTokens":0,
    "inputTokens":8,"outputTokens":20,"serverToolUsage":{},"totalTokens":28
  }
}

## Summary Reverts NVIDIA#963 and adds a guard test that prevents new CSP-specific provider keys from being added to the installer. ## Rationale Bedrock's OpenAI-compatible endpoint already works through the existing `compatible-endpoint` option — the same path every other CSP wrapper (Vertex, Azure OpenAI, etc.) should use. Adding a dedicated provider key for each CSP does not scale: each one requires config in `onboard.js`, a case in `inference-config.js`, UX menu entries, model curation, validation logic, tests, and docs. The `compatible-endpoint` and `compatible-anthropic-endpoint` options already cover these use cases generically. ## Changes - **`bin/lib/onboard.js`**: Remove `BEDROCK_ENDPOINT_URL`, `bedrock` from `REMOTE_PROVIDER_CONFIG`, `REMOTE_MODEL_OPTIONS`, `validProviders`, interactive menu, region check, validation, `setupInference`, and dashboard label. - **`bin/lib/inference-config.js`**: Remove `bedrock` case from `getProviderSelectionConfig()`. - **`test/inference-config.test.js`**: Remove Bedrock-specific assertions; add `does not grow beyond the approved provider set` guard test that fails if any new provider key (bedrock, vertex, azure, etc.) is introduced. ## Related - Reverts NVIDIA#963 - Closes NVIDIA#1204 ## Type of Change - [x] Code change for a new feature, bug fix, or refactor. ## Checklist - [x] Formatters applied - [x] Tests added (guard test) - [x] No secrets committed - [x] `npm test` passes (725 passed)  ## Summary by CodeRabbit * **Breaking Changes** * Removed support for Amazon Bedrock as an inference provider. * **Documentation** * Updated command documentation to recommend `nemoclaw onboard` for new installations. * Marked legacy `nemoclaw setup` command as deprecated. * Updated deployment guides to reference the `nemoclaw onboard` setup wizard. * **Tests** * Updated test suite to validate approved inference providers.  Co-authored-by: Deepak Jain <deepujain@users.noreply.github.com>

cv mentioned this pull request Mar 31, 2026

test: add unit tests for Bedrock provider #1204

Closed

13 tasks

Merge branch 'main' into revert/bedrock-provider

8a1a058

cv enabled auto-merge (squash) March 31, 2026 22:47

kjw3 approved these changes Mar 31, 2026

View reviewed changes

cv merged commit 182ad1c into main Mar 31, 2026
3 of 4 checks passed

cv deleted the revert/bedrock-provider branch March 31, 2026 22:48

wscurran added the refactor PR restructures code without intended behavior change label Jun 8, 2026

Navigation Menu

Uh oh!

refactor: remove Amazon Bedrock as first-class provider - #1212

refactor: remove Amazon Bedrock as first-class provider#1212
cv merged 2 commits into
mainfrom
revert/bedrock-provider

cv commented Mar 31, 2026 •

edited by coderabbitai Bot

Loading

Uh oh!

coderabbitai Bot commented Mar 31, 2026 •

edited

Loading

Review failed

Walkthrough

Changes

Estimated code review effort

Possibly related PRs

Suggested reviewers

Poem

❌ Failed checks (2 warnings)

Uh oh!

github-actions Bot commented Mar 31, 2026

Uh oh!

Uh oh!

VKudlay commented Apr 2, 2026 •

edited

Loading

Uh oh!

VKudlay commented Apr 2, 2026 •

edited

Loading

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Uh oh!

Conversation

cv commented Mar 31, 2026 • edited by coderabbitai Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Rationale

Changes

Related

Type of Change

Checklist

Summary by CodeRabbit

Release Notes

Uh oh!

coderabbitai Bot commented Mar 31, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Review failed

Walkthrough

Changes

Estimated code review effort

Possibly related PRs

Suggested reviewers

Poem

❌ Failed checks (2 warnings)

Uh oh!

github-actions Bot commented Mar 31, 2026

Uh oh!

Uh oh!

VKudlay commented Apr 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

VKudlay commented Apr 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

cv commented Mar 31, 2026 •

edited by coderabbitai Bot

Loading

coderabbitai Bot commented Mar 31, 2026 •

edited

Loading

VKudlay commented Apr 2, 2026 •

edited

Loading

VKudlay commented Apr 2, 2026 •

edited

Loading