4 Commits

Author SHA1 Message Date
Pietro Zullo 8d4cb4719b Feature/mcp 799 python client integration of code mode (#498)
* initial commit

* update to code mode

* Squash all commits: Add CLI deploy functionality, server tool inspector, and various improvements (#467)

* chore(release): exit prerelease mode and version packages (#473)

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

* Docs/tunnel docs (#474)

* chore: add tunnelling docs

* chore: add tunnelling docs

* docs: enhance tunneling documentation with link to @mcp-use/tunnel package

* docs: update tunneling documentation with improved code examples and formatting

* feat: Add verify parameter to MCPClient and connectors (#478)

* feat: add verify parameter to MCPClient and connectors

- Introduced a new optional `verify` parameter in `MCPClient`, `HttpConnector`, and configuration functions to control SSL certificate verification.
- Updated relevant methods to utilize the `verify` parameter, ensuring consistent handling of SSL verification across the client and connectors.
- Enhanced the SSE and Streamable HTTP connection managers to accept a custom HTTPX client factory for improved flexibility in connection handling.

* fix: linting and docs

* fix: http connector test

* Update docs/python/api-reference/mcp_use_client_config.mdx

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Signed-off-by: Pietro Zullo <62951181+pietrozullo@users.noreply.github.com>

* fix: docs

---------

Signed-off-by: Pietro Zullo <62951181+pietrozullo@users.noreply.github.com>
Co-authored-by: Pietro Zullo <62951181+pietrozullo@users.noreply.github.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Canary release (#476)

* Implement iframe log console for inspector (#475)

* Docs/tunnel docs (#474)

* chore: add tunnelling docs

* chore: add tunnelling docs

* docs: enhance tunneling documentation with link to @mcp-use/tunnel package

* docs: update tunneling documentation with improved code examples and formatting

* feat: Add iframe console logging and UI

Implement a system to capture and display console logs from iframes. This includes injecting a script into iframes to intercept console messages and a UI component to display them in the parent window.

Co-authored-by: tonxipad <tonxipad@gmail.com>

* feat: Conditionally inject console interceptor for same-origin iframes

Co-authored-by: tonxipad <tonxipad@gmail.com>

* refactor: update IframeConsole component for improved UI and styling

- Removed unused XIcon import.
- Adjusted LogLevelBadge styling for better visual consistency.
- Renamed "Iframe Console" to "Widget Console Logs" for clarity.
- Modified button styling for better alignment and spacing.

* feat: add HTMLIFrameElement and MessageEvent to ESLint config

- Updated ESLint configuration to include HTMLIFrameElement and MessageEvent as readonly types, enhancing type safety for iframe-related code.

* feat: implement auto-scroll functionality in IframeConsole component

- Added useRef and useEffect hooks to enable auto-scrolling to the bottom of the console when logs change or the console opens.
- Updated the console's container div to use the scrollContainerRef for managing scroll position.

* chore: lint & format

* chore: changeset

---------

Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: tonxipad <tonxipad@gmail.com>

* chore: enter canary prerelease mode

* chore(typescript): version packages (canary)

* Feat/consistent sub for tunnel (#480)

* Docs/tunnel docs (#474)

* chore: add tunnelling docs

* chore: add tunnelling docs

* docs: enhance tunneling documentation with link to @mcp-use/tunnel package

* docs: update tunneling documentation with improved code examples and formatting

* Refactor CLI tunnel functionality to support optional subdomain and save to manifest

- Updated `startTunnel` function to accept an optional subdomain parameter.
- Enhanced subdomain extraction from the tunnel URL and updated the manifest file with the subdomain.
- Cleaned up commented-out code related to tunnel handling in the CLI.
- Improved logging for tunnel establishment and subdomain saving.

* chore: changeset

* format

* chore(typescript): version packages (canary)

* Refactor tunnel argument handling in CLI to streamline subdomain flag usage and improve logging

* remove console log

* [typescript][server] Fix tools with async calls (#481)

* fix tools with async calls

* FORMAT

* chore(typescript): version packages (canary)

* fix/minor fix post canary (#482)

* fix server async call

* wrap all handle request calls in wait function

* changeset

* chore(typescript): version packages (canary)

* window.openai emulation and ChatGPT Integration debugging (#469)

* feat: Add WidgetFullscreenWrapper component

Co-authored-by: tonxipad <tonxipad@gmail.com>

* Docs/tunnel docs (#474)

* chore: add tunnelling docs

* chore: add tunnelling docs

* docs: enhance tunneling documentation with link to @mcp-use/tunnel package

* docs: update tunneling documentation with improved code examples and formatting

* feat: Enhance OpenAIComponentRenderer with tool call handling and display mode management

- Implemented tool call handling with error management and response formatting for OpenAI API compatibility.
- Added support for follow-up messages and localStorage fallback for pending follow-ups.
- Introduced display mode management (inline, pip, fullscreen) with dynamic iframe resizing and a close button for fullscreen mode.
- Updated TypeScript configuration to include .tsx files for better component support.

* fix: maintain subdomain across builds

* remove console

* refactor: replace direct serverBaseUrl usage with getServerBaseUrl method for improved consistency

* refactor: enhance WidgetFullscreenWrapper with hoverable icon buttons for fullscreen and pip modes

* feat: implement dev widget proxy and asset handling for improved development experience

* refactor: update OpenAIComponentRenderer to handle global updates for display mode and theme, enhancing iframe communication

* feat: add WidgetDebugger export to enhance debugging capabilities in the React package

* chore: fix type

* format

* chore: changeset

* remove irrelevant content

* chore: add changeset for MCP-747-fix-apps-sdk-fullscreen-close-button-9e35

* refactor: improve tunnel shutdown handling and output suppression for better user experience

* format

* remove console log

---------

Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: tonxipad <tonxipad@gmail.com>

* chore(typescript): version packages (canary)

* Fix/settings button servers dropdown (#485)

* feat: Add server connection management in Layout and ServerDropdown components

- Introduced ServerConnectionModal for editing connection settings in the Layout component.
- Updated ServerDropdown to handle connection options and display selected server status.
- Enhanced MCPConnection interface to include transportType and proxyConfig for better connection management.
- Implemented connection update logic with success notifications.

* add gear also to home

* chore: changeset and format

* chore(typescript): version packages (canary)

* feat: Enhance ConnectionSettingsForm layout and add utility styles (#486)

- Updated ConnectionSettingsForm component to improve layout responsiveness with new utility classes.
- Added a new CSS utility for flex column and row layout.
- Cleaned up imports for better organization.

* chore(typescript): version packages (canary)

* Update libraries/typescript/packages/cli/CHANGELOG.md

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Signed-off-by: Enrico Toniato <2827496+tonxxd@users.noreply.github.com>

* Update libraries/typescript/packages/cli/src/index.ts

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Signed-off-by: Enrico Toniato <2827496+tonxxd@users.noreply.github.com>

* Update libraries/typescript/packages/cli/src/index.ts

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Signed-off-by: Enrico Toniato <2827496+tonxxd@users.noreply.github.com>

* Update libraries/typescript/.changeset/clear-wombats-juggle.md

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Signed-off-by: Enrico Toniato <2827496+tonxxd@users.noreply.github.com>

* Feature/mcp 743 tool argument autoformats and moves the cursor (#487)

* refactor(ToolsTab, ToolInputForm): enhance JSON handling for object/array types

- Updated ToolsTab to initialize object and array types as empty JSON strings to preserve formatting.
- Improved argument parsing in ToolsTab to handle JSON strings for object/array types before tool execution.
- Modified ToolInputForm to display objects/arrays as formatted JSON strings, ensuring user input is preserved.

* chore: changeset

* chore(typescript): version packages (canary)

* Fix/remove console button from chat (#488)

* remove mcp url console log

* feat(OpenAIComponentRenderer): add showConsole prop to control console visibility

* Feature/mcp 837 load list of models for providers (#489)

* remove mcp url console log

* Enhance ConfigurationDialog with model fetching and caching

- Added functionality to fetch models from OpenAI, Anthropic, and Google APIs based on the selected provider.
- Implemented caching for fetched models using localStorage to improve performance and reduce API calls.
- Updated the UI to display loading states and errors when fetching models.
- Refactored API key handling to support provider-specific storage in localStorage.
- Improved user experience by debouncing API calls and dynamically updating the model selection based on the provider.

* fix: reorder imports in useConfig.ts for better clarity

* chore(typescript): version packages (canary)

---------

Signed-off-by: Enrico Toniato <2827496+tonxxd@users.noreply.github.com>
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: tonxipad <tonxipad@gmail.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Pietro Zullo <62951181+pietrozullo@users.noreply.github.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* chore(release): exit prerelease mode and version packages (#490)

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

* Update Dockerfile

Signed-off-by: Enrico Toniato <2827496+tonxxd@users.noreply.github.com>

* fix: x icon light mode

* docs: Improve typescript server docs (#491)

* fix: update installation commands and enhance quickstart documentation

- Changed `pnpm run dev` to `pnpm dev` in installation instructions across multiple files for consistency.
- Improved the quickstart guide by clarifying the MCP server framework's capabilities and updating the project structure details.
- Added a base URL configuration for the MCP server in the configuration documentation to ensure proper functionality of UI widgets.

* minor

* add fallback

---------

Signed-off-by: Luigi Pederzani <luigipederzani@gmail.com>

* Docs/inspector (#492)

* refactor: reorganize documentation structure for Inspector and Dev Tools

- Updated the docs.json file to consolidate Inspector and Tunneling pages under a new "Dev Tools" group.
- Removed redundant entries for Inspector and Tunneling from the documentation structure.
- Added a sidebar title for the Inspector page to enhance navigation.

* docs: enhance Inspector documentation and add new CLI usage guides

- Updated `docs.json` to include new pages for CLI usage, connection settings, keyboard shortcuts, command palette, integration, debugging ChatGPT apps, and development under the Inspector section.
- Expanded the Inspector documentation to cover new features, including CLI usage and detailed guides for keyboard shortcuts and integration.
- Improved the `self-hosting` documentation to reference the new CLI usage guide.
- Enhanced the `apps-sdk-resources` documentation with additional details on widget state and testing procedures using the Inspector.

* fix broken links

* chore: remove unuseful docs

* Feat/python/pretty display (#497)

* pretty display implementation

* docs update

* initial commit

* docs

* fix tests

* example streams and uses anthropic

* docs

* docs image

* fix type checking mock

* add rich to requirements

* enable pretty pritns in all examples

---------

Signed-off-by: Pietro Zullo <62951181+pietrozullo@users.noreply.github.com>
Signed-off-by: Enrico Toniato <2827496+tonxxd@users.noreply.github.com>
Signed-off-by: Luigi Pederzani <luigipederzani@gmail.com>
Co-authored-by: Enrico Toniato <2827496+tonxxd@users.noreply.github.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Vincenzo Reina <vincenzoreinaa05@gmail.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: tonxipad <tonxipad@gmail.com>
Co-authored-by: Luigi Pederzani <luigipederzani@gmail.com>
2025-11-20 16:26:23 +01:00
Luigi Pederzani f08e870e09 feat: Upgrade to Langchain 1.0.0 (#393)
* chore: update imports

* refactor(mcpagent): update agent creation logic and imports

- Replaced `create_tool_calling_agent` with `create_agent` for improved agent instantiation.
- Simplified the `_create_agent` method by removing the `AgentExecutor` and directly returning the created agent.
- Adjusted type annotations for `_agent_executor` to initialize as `None` instead of using a union type.

* feat(mcpagent): migrate to LangChain 1.0.0 and simplify agent methods

- Updated MCPAgent to utilize LangChain 1.0.0's `create_agent()` for improved agent instantiation.
- Introduced new methods `astream_simplified()` and `run_v2()` that leverage the built-in `astream()` for internal agent loop handling.
- Retained legacy methods `stream()` and `run()` for backward compatibility with manual execution.
- Enhanced documentation to reflect changes and provide clearer usage examples.

* feat(mcpagent): implement dynamic tool updates during agent execution

- Enhanced the MCPAgent to support dynamic tool updates when using server_manager mode.
- Implemented logic to check for tool changes before and during execution, allowing for immediate updates and safe restarts.
- Introduced a maximum restart limit to prevent infinite loops during tool updates.
- Updated documentation to reflect the new behavior and usage of dynamic tool management.

* refactor(mcpagent): clean up imports and streamline logging

- Removed unused imports and organized existing ones for better clarity in `mcpagent.py`.
- Simplified logging statements for consistency and improved readability.
- Enhanced the overall structure of the code by eliminating unnecessary whitespace.
- Updated unit tests to reflect changes in imports and ensure compatibility with the refactored code.

* feat(mcpagent): enhance agent execution with step tracking and middleware

- Introduced `ModelCallLimitMiddleware` to enforce a maximum number of model calls during agent execution.
- Updated the `_consume_and_return` method to return a tuple of the final result and the number of steps taken.
- Enhanced telemetry tracking to include steps taken and execution time for better performance insights.
- Simplified agent creation logic by integrating middleware directly into the `create_agent` function.
- Improved logging for agent execution details, including steps and tool usage.

* fix(mcpagent): add null check for node output messages extraction

- Updated the MCPAgent to include a null check for `node_output` before attempting to access the "messages" key, preventing potential errors during message extraction.

* minor

* feat(mcpagent): filter out ToolMessage from history

- Updated the MCPAgent to exclude ToolMessage instances from being added to the message history, ensuring that only relevant messages are recorded.
- This change helps maintain a cleaner history by avoiding intermediate tool execution details.

* fix(mcpagent): set max_steps assignment

* ruff

* docs

* e2e tests for agent

* key in tests

* fix manager test and split ci tests

* telemetry off for tests

* fix server manager test

* feat(mcpagent): delegate query streaming to remote agent in remote mode

- recreated functionality to the MCPAgent to yield results from a remote agent when operating in remote mode, enhancing the agent's capability to handle queries efficiently.

* fix: agent unit tests

* Print inputs

* rollback, wrong repo

* fix streaming output

* stricter streaming test

* pass observability callbacks

* fix tests mock

---------

Co-authored-by: pietrozullo <pietro.zullo@gmail.com>
2025-10-27 14:26:01 +01:00
Pietro Zullo 0f60da81a4 Feat: structured output (#203)
* mcpagent structured output

* Structured output example

* docs

* fix stream example and ensure it works
2025-07-31 21:48:59 -07:00
Pietro Zullo c789c317a7 Feat: streaming steps (#159)
* Streaming steps implementation and example

* Changed astream to stream_events for clarity

* docs

* Changed file docsg

* Fixed telemetry
2025-07-05 17:22:31 -07:00