> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mcp-assistant.in/llms.txt
> Use this file to discover all available pages before exploring further.

# April 21, 2026

> Release notes for mcp-ts v1.5.0: the new MCP Gateway, Tool Router middleware, AI framework adapters, Supabase storage, and additional improvements.

## New features

### Tool Router

You can now intelligently manage large tool catalogs with the new [Tool Router](/core-concepts/tool-router). Instead of injecting every tool schema into the LLM's context window, the Tool Router lets you choose from three strategies:

* **`all`** — Pass all tools through (the default, great for small projects).
* **`search`** — Replace your tools with four lightweight [Meta-Tools](/core-concepts/meta-tools) that let the LLM discover, inspect, and execute tools on demand. This scales to 1,000+ tools with minimal token usage.
* **`groups`** — Expose specific sets of tools based on application state or user intent.

The `search` strategy supports both keyword-based BM25 matching and optional semantic search via a custom embedding function.

### MCP Gateway

The new [MCP Gateway](/gateway/overview) lets you expose local MCP servers to cloud-based AI clients like ChatGPT and Claude. It creates a secure tunnel so remote assistants can interact with tools running on your machine — no open ports or firewall changes required.

* Install and run via `uvx mcpassistant-gateway`
* Configure servers through a single [`mcp.json`](/gateway/configuration) file
* Manage all local-to-cloud connections through one persistent bridge

### AI framework adapters

Four new [adapters](/ai-adapters/overview) make it easy to use MCP tools with popular AI frameworks:

* [**AI SDK**](/ai-adapters/ai-sdk) — Drop MCP tools into any Vercel AI SDK project.
* [**LangChain**](/ai-adapters/langchain) — Use MCP tools as LangChain-compatible tools in chains and agents.
* [**Mastra**](/ai-adapters/mastra) — Integrate MCP tools with the Mastra agent framework.
* [**AG-UI**](/ai-adapters/agui) — Connect MCP tools to remote agents via the AG-UI protocol, with built-in middleware for server-side tool execution.

### Supabase storage backend

[Supabase](/storage-backends/supabase) is now available as a production-ready storage backend. It includes:

* PostgreSQL persistence with JSONB support
* Row-level security (RLS) for tenant isolation
* Application-level **AES-256-GCM encryption** for sensitive session fields like tokens and headers
* Auto-detection via `SUPABASE_URL` environment variable
* A built-in `supabase-init` CLI command to scaffold the required migration files

### MCP Documentation Server

The MCP Toolkit documentation is now available as a hosted [MCP server](/modelcontextprotocol). You can add it to any MCP-compatible client — such as Claude Desktop, Cursor, or VS Code — to give your AI assistant direct access to search, browse, and read the full docs. No more copy-pasting; your tools stay in sync with the latest documentation automatically.

### LocalStorage backend

A new [LocalStorage backend](/storage-backends/overview) lets you persist sessions directly in the browser. This is ideal for client-side apps and prototyping where you don't need a server-side database. It supports configurable namespaces, TTL-based expiration, and lazy eviction — all scoped to the current origin.

### MCP App Host integration

MCP Apps now support a full [App Host](/mcp-apps) integration for rendering interactive tool UIs in sandboxed iframes. Key additions:

* **Sandbox proxy** with configurable Content Security Policy (CSP)
* **Resource preloading** — UI resources are cached when tools are discovered, so apps open faster
* **Fullscreen mode** — Guest apps can request fullscreen rendering via the AppBridge protocol
* **Host ↔ guest sync** — Streaming partial input, cancellation, and host context (theme, display mode) are all synchronized automatically
* **Custom `loader` prop** — Show your own loading UI while an MCP App initializes

## Updates

### Project rebrand to MCP Toolkit

The project has been renamed from "Toolkit" to **MCP Toolkit** to better reflect its purpose as a comprehensive toolkit for the Model Context Protocol.

### New documentation site

All documentation has moved to a new [Mintlify-powered site](https://docs.mcp-assistant.in). The new site features improved navigation, search, and a refreshed layout. Storage backend and AI adapter pages now live under dedicated sections, and the API reference has been consolidated into a single tab.

### Simplified MCP App rendering

You can now import `McpAppRenderer` and `getMcpAppMetadata` directly from the [React client](/react) instead of going through the `useMcpApps` hook. This gives you more control over when and where MCP App UIs are rendered. The previous hook-based API still works but is deprecated.

### Adapter namespace support

Tool adapters now support namespacing, which prevents name collisions when connecting to multiple MCP servers. Set a custom prefix per adapter to keep tool names unique.

### Isomorphic storage architecture

Storage backends now work across both server and client environments. You can inject a storage backend directly into the MCP client or multi-session client via a `storage` option, making it easy to swap between server-side backends (Redis, Supabase) and client-side backends (LocalStorage) without changing your application code.

### Gemini CLI compatibility

The server now correctly responds to Gemini CLI's `streamable-http` transport checks and MCP ping requests, so you can use MCP Toolkit with Google's Gemini CLI without additional configuration.

### Next.js example refresh

The [Next.js guide](/nextjs) now includes an updated full-stack example with App Router support, an integrated chat UI, and inline MCP App rendering after tool calls.

## Bug fixes

* **OAuth handling** — Improved OAuth flow reliability in the `useMcp` hook, including better callback handling and storage of OAuth credentials.
* **MCP Apps renderer stability** — Fixed iframe white-screen flashes on load and stabilized client identity handling in the React renderer.
* **Tool name length violations** — Resolved an issue where generated tool names could exceed protocol limits, causing connection failures.
* **Event routing** — Fixed shared event routing so connection state updates are correctly reflected in the UI.
* **Storage proxy deadlock** — Fixed a deadlock in the storage proxy that could cause hangs during backend initialization.
