feat: add NVIDIA NIM as an LLM provider

[Jon Patterson (freshyjmp)]

What this changes

Adds NVIDIA NIM as a built-in LLM provider (nvidia). Users can pick an NVIDIA NIM card in AI settings, enter an API key, and run the BrowserOS agent on NVIDIA-hosted models through NIM's OpenAI-compatible API. The base URL is pre-filled with https://integrate.api.nvidia.com/v1 and can be changed to point at a self-hosted NIM.

• packages/browseros-agent/packages/shared/src/schemas/llm.ts: adds LLM_PROVIDERS.NVIDIA and adds it to LLMProviderSchema.
• packages/browseros-agent/apps/server/src/agent/provider-factory.ts, packages/browseros-agent/apps/server/src/lib/clients/llm/provider.ts: add nvidia factories built on createOpenAICompatible, following the Moonshot provider (base URL and API key are required, custom headers are forwarded).
• packages/browseros-agent/apps/app/lib/llm-providers/{types.ts,providerTemplates.ts,providerIcons.tsx}, packages/browseros-agent/apps/app/screens/ai-settings/{add-provider.helpers.ts,provider-form-schema.ts}: add the NVIDIA NIM template (default model nvidia/nemotron-3-super-120b-a12b, text only), its icon, and form support.
• packages/browseros-agent/apps/server/tests/agent/provider-factory-nvidia.test.ts: new tests for the factory (base URL, API key, missing credentials, headers).
• docs/browseros/features/bring-your-own-llm.mdx: adds an NVIDIA NIM section.

No dependencies were added.

Why

NIM serves NVIDIA-hosted models over an OpenAI-compatible API. Before this change, users had to set it up by hand as a generic OpenAI-compatible provider. A dedicated provider gives them a ready-made template with the correct endpoint and a default model that supports tool calls, which the agent loop needs.

How to check it

Path: BrowserOS (side panel chat / agent loop).

1. Go to chrome://browseros/settings.
2. Click USE on the NVIDIA NIM card.
3. Enter an NVIDIA API key from build.nvidia.com. The model is pre-set to nvidia/nemotron-3-super-120b-a12b and the base URL to https://integrate.api.nvidia.com/v1.
4. Set Context Window, click Save, then chat with the agent and run a task that uses tools.

Checks run:

• cd packages/browseros-agent/apps/server && bun --env-file=../../.env.development test ./tests/agent/provider-factory-nvidia.test.ts: 8 pass, 0 fail
• cd packages/browseros-agent/apps/server && bun --env-file=../../.env.development test ./tests/agent/provider-factory-baseurl.test.ts: 18 pass, 0 fail
• cd packages/browseros-agent/apps/app && bun run ../../scripts/run-bun-test.ts --cwd=apps/app ./apps/app/lib/llm-providers/providerTemplates.test.ts ./apps/app/screens/ai-settings/add-provider.helpers.test.ts: pass
• bunx @biomejs/biome check on the changed TypeScript files: no issues
• bun run typecheck in apps/server and apps/app: pass
• git diff --check -- docs/browseros/features/bring-your-own-llm.mdx: clean
• Live check against https://integrate.api.nvidia.com/v1 with nvidia/nemotron-3-super-120b-a12b, using the server's createLLMProvider: a chat reply and a tool call both succeeded

Before you request review

• Ran the checks for the path you touched (see CONTRIBUTING.md)
• Added or updated tests where the behaviour changed
• Checked this on the platforms it affects

Validation

Independent validation: passed on a clean checkout of b768a3f68ba9 with this patch applied.

• tool-setup-1: passed
• setup-install: passed
• setup-env: passed
• test-nvidia-factory: passed
• test-peer-factory: passed
• test-app-providers: passed
• docs-diff-check: passed
• lint: passed
• typecheck-server: passed
• typecheck-app: passed
• nim-smoke: passed
• nim-smoke-result: passed

[github-actions]

Thank you for your contribution! Before we can merge this PR, we need you to sign our Contributor License Agreement.

To sign the CLA, please add a comment to this PR with the following text:

I have read the CLA Document and I hereby sign the CLA

You only need to sign once. After signing, this check will pass automatically.

---

Troubleshooting

• Already signed but still failing? Comment recheck to trigger a re-verification.
• Signed with a different email? Make sure your commit email matches your GitHub account email, or add your commit email to your GitHub account.

- - - I have read the CLA Document and I hereby sign the CLA - - - **NIM OSS Manager** seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please [add the email address used for this commit to your account](https://help.github.com/articles/why-are-my-commits-linked-to-the-wrong-user/#commits-are-not-linked-to-any-user).
_{You can retrigger this bot by commenting **recheck** in this Pull Request. Posted by the **CLA Assistant Lite bot**.}