[DOCS] Add NVIDIA provider to Rust ADK provider docs

[inference-gateway-maintainer]

The Rust ADK (inference-gateway/rust-adk) now supports NVIDIA as a first-class LLM provider, propagating the ecosystem-wide provider addition (gateway -> schemas -> SDKs -> ADKs).

What changed (user-facing):

• A2A_AGENT_CLIENT_PROVIDER=nvidia is now accepted by the Rust ADK (mapped to the SDK Provider::Nvidia variant after bumping inference-gateway-sdk to 0.14.1).
• Example model id: nvidia/meta/llama-3.1-8b-instruct (NVIDIA NIM via the build.nvidia.com API catalog).
• NVIDIA_API_KEY was added to the example .env files and to the inference-gateway service passthrough in the example docker-compose files.

Docs to update:

• Any provider matrix / supported-providers list that enumerates Rust ADK (or ecosystem) providers should include NVIDIA.
• If there is a per-provider setup page, add NVIDIA (bearer auth, base URL https://integrate.api.nvidia.com/v1, supports_vision: true).

Reference PR: inference-gateway/rust-adk#86 (Closes inference-gateway/rust-adk#84).

[edenreich]

@claude Can you work on this?

[inference-gateway-maintainer]

Claude finished @edenreich's task in 9m 37s —— View job • docs/issue-343-20260706-0001 • [Create PR ➔](https://github.com/inference-gateway/docs/compare/main...docs/issue-343-20260706-0001?quick_pull=1&title=Issue%20%23343%3A%20Changes%20from%20Claude&body=This%20PR%20addresses%20issue%20%23343%0A%0AGenerated%20with%20%5BClaude%20Code%5D(https%3A%2F%2Fclaude.ai%2Fcode)

---

Done: NVIDIA added to the Rust ADK provider docs

• Gather context (read rust-adk.md, provider docs, generator setup)
• Understand what needs updating for NVIDIA
• Add NVIDIA to the Rust ADK provider docs (rust-adk.md)
• Run lint / format / build / tests
• Commit and open PR

What I found

NVIDIA was already fully documented at the gateway level - supported-providers.md (provider matrix, vision list, request example), configuration.md (settings + .env), and the generator fixture/overrides all include it. So the "provider matrix / per-provider setup page" requirements from the issue were already satisfied there.

The real gap was the Rust ADK page (rust-adk.md), which had no provider enumeration at all - just (e.g. openai, ollama, groq) on the A2A_AGENT_CLIENT_PROVIDER row, unlike the sibling TypeScript ADK page which has a dedicated provider table.

What I changed (rust-adk.md)

• Added a "Switching providers and models" subsection under AgentBuilder, mirroring the TypeScript ADK page's table, including NVIDIA: nvidia / nvidia/meta/llama-3.1-8b-instruct / NVIDIA_API_KEY, bearer auth at https://integrate.api.nvidia.com/v1 (build.nvidia.com NIM catalog).
• Added nvidia to the A2A_AGENT_CLIENT_PROVIDER env-var reference examples.

Verification

bun run format:check, bun run lint:md, bun run build, and bun test (6 pass) all succeed.

Note / possible follow-up

The example .env and docker-compose changes mentioned in the issue live in the rust-adk repo (already shipped in rust-adk#86), not this docs repo. Separately, the sibling TypeScript ADK provider table (typescript-adk.md) also predates NVIDIA - aligning it would be a reasonable follow-up, but I scoped this to the Rust ADK per the issue.

PR: #345

• branch: docs/issue-343-20260706-0001

[inference-gateway-releaser]

🎉 This issue has been resolved in version 0.9.0 🎉

The release is available on GitHub release

Your semantic-release bot 📦🚀