feat(ai): modernize providers and add LiteLLM/Ollama adapters

[willgriffin]

What changed

This PR modernizes @happyvertical/ai as a broader provider cleanup release rather than a single-provider patch.

It adds first-class litellm and ollama adapters, updates provider and auto-detection wiring, improves model discovery and capability reporting, and fills in several previously thin or misleading provider paths.

It also expands integration coverage so we can validate real gateways and local model hosts instead of relying only on mocked behavior.

Highlights

• add a first-class LiteLLM provider with provider-specific attribution, model resolution, and live integration coverage against llm.happyvertical.com
• add a first-class Ollama provider with native chat, streaming, embeddings, vision description, image embedding fallback, and OpenAI-compatible image generation bridging
• modernize Gemini and Bedrock provider implementations and tighten provider capability reporting across the package
• improve Node-side provider auto-detection for LiteLLM, Ollama, Anthropic, and Gemini environments
• update package docs, metadata, package surface, and generated agent documentation to reflect the expanded provider set
• add deeper provider and integration tests, including live Ollama validation against internal nodes

Why

We want to make heavy use of LiteLLM and local Ollama infrastructure, and the package had started to accumulate a mix of OpenAI-shaped assumptions, partial provider implementations, and uneven integration coverage.

This release closes those gaps so the package is a better fit for heterogeneous model routing, local model hosts, and current SDK behavior.

Impact

Consumers can now configure type: 'litellm' or type: 'ollama' directly instead of routing everything through the OpenAI adapter.

Provider selection, capability reporting, model discovery, and observability are more honest across the supported backends, and the package now has stronger live validation for the environments we actually run.

Validation

• pnpm --filter @happyvertical/ai build
• set -a && source .env.local && set +a && pnpm --filter @happyvertical/ai test
• OLLAMA_HOST=snail-ollama:11434 OLLAMA_INTEGRATION=1 pnpm --filter @happyvertical/ai test -- src/ollama.integration.test.ts
• OLLAMA_HOST=warthog-ollama:11434 OLLAMA_INTEGRATION=1 pnpm --filter @happyvertical/ai test -- src/ollama.integration.test.ts
• pre-push hook: turbo run typecheck

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: bc19b6be2b

ℹ️ About Codex in GitHub

Your team has set up Codex to 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 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[github-actions]

📦 Version Bump Preview

When this PR is merged, packages will receive a minor
version bump based on your conventional commits.

What happens on merge?

1. Tests run on main branch
2. Packages are built
3. Versions are bumped automatically
4. Packages are published to GitHub Packages
5. Git tags are created

No manual intervention needed! 🎉