Add support for MCP servers

frontend/src/i18n/declaration.ts

@@ -289,6 +289,7 @@ export enum I18nKey {
   OBSERVATION_MESSAGE$EDIT = "OBSERVATION_MESSAGE$EDIT",
   OBSERVATION_MESSAGE$WRITE = "OBSERVATION_MESSAGE$WRITE",
   OBSERVATION_MESSAGE$BROWSE = "OBSERVATION_MESSAGE$BROWSE",
+  OBSERVATION_MESSAGE$MCP_CALL_TOOL = "OBSERVATION_MESSAGE$MCP_CALL_TOOL",
   EXPANDABLE_MESSAGE$SHOW_DETAILS = "EXPANDABLE_MESSAGE$SHOW_DETAILS",
   EXPANDABLE_MESSAGE$HIDE_DETAILS = "EXPANDABLE_MESSAGE$HIDE_DETAILS",
   AI_SETTINGS$TITLE = "AI_SETTINGS$TITLE",

frontend/src/i18n/translation.json

@@ -4313,6 +4313,21 @@
         "es": "Navegación completada",
         "tr": "Gezinme tamamlandı"
     },
+    "OBSERVATION_MESSAGE$MCP_CALL_TOOL": {
+        "en": "Called an MCP tool",
+        "zh-CN": "调用了 MCP 工具",
+        "zh-TW": "調用了 MCP 工具",
+        "ko-KR": "MCP 도구 호출됨",
+        "ja": "MCPツールを呼び出しました",
+        "no": "Kalt et MCP-verktøy",
+        "ar": "تم استدعاء أداة MCP",
+        "de": "Ein MCP-Tool aufgerufen",
+        "fr": "Outil MCP appelé",
+        "it": "Chiamato uno strumento MCP",
+        "pt": "Ferramenta MCP chamada",
+        "es": "Se ha llamado a una herramienta MCP",
+        "tr": "Bir MCP aracı çağrıldı"
+    },
     "EXPANDABLE_MESSAGE$SHOW_DETAILS": {
         "en": "Show details",
         "zh-CN": "显示详情",

frontend/src/services/observations.ts

@@ -50,6 +50,7 @@ export function handleObservationMessage(message: ObservationMessage) {
     case ObservationType.READ:
     case ObservationType.EDIT:
     case ObservationType.THINK:
+    case ObservationType.MCP_CALL_TOOL:
     case ObservationType.NULL:
       break; // We don't display the default message for these observations
     default:
@@ -125,6 +126,21 @@ export function handleObservationMessage(message: ObservationMessage) {
           }),
         );
         break;
+      case "mcp_call_tool":
+        store.dispatch(
+          addAssistantObservation({
+            ...baseObservation,
+            observation: "mcp_call_tool" as const,
+            extras: {
+              tool_name: String(message.extras.tool_name || ""),
+              kwargs:
+                typeof message.extras.kwargs === "object"
+                  ? (message.extras.kwargs as Record<string, unknown>)
+                  : {},
+            },
+          }),
+        );
+        break;
       case "delegate":
         store.dispatch(
           addAssistantObservation({

frontend/src/state/chat-slice.ts

@@ -22,6 +22,7 @@ const HANDLED_ACTIONS: OpenHandsEventType[] = [
   "browse",
   "browse_interactive",
   "edit",
+  "mcp_call_tool",
 ];
 
 function getRiskText(risk: ActionSecurityRisk) {
@@ -112,6 +113,12 @@ export const chatSlice = createSlice({
       } else if (actionID === "browse_interactive") {
         // Include the browser_actions in the content
         text = `**Action:**\n\n\`\`\`python\n${action.payload.args.browser_actions}\n\`\`\``;
+      } else if (actionID === "mcp_call_tool") {
+        text = `Tool name: \`${action.payload.args.tool_name}\`\n\nArguments:\n\`\`\`python\n${JSON.stringify(
+          action.payload.args.kwargs,
+          null,
+          2,
+        )}\n\`\`\``;
       }
       if (actionID === "run" || actionID === "run_ipython") {
         if (
@@ -193,6 +200,18 @@ export const chatSlice = createSlice({
         } else {
           causeMessage.content = observation.payload.content;
         }
+      } else if (observationID === "mcp_call_tool") {
+        let { content } = observation.payload;
+        if (content.length > MAX_CONTENT_LENGTH) {
+          content = `${content.slice(0, MAX_CONTENT_LENGTH)}...`;
+        }
+        const toolName = observation.payload.extras.tool_name;
+        const toolKwargs = JSON.stringify(
+          observation.payload.extras.kwargs,
+          null,
+          2,
+        );
+        causeMessage.content = `Tool name: \`${toolName}\`\n\nArguments:\n\`\`\`python\n${toolKwargs}\n\`\`\`\n\nOutput:\n\`\`\`python\n${content.trim()}\n\`\`\``;
       } else if (observationID === "browse") {
         let content = `**URL:** ${observation.payload.extras.url}\n`;
         if (observation.payload.extras.error) {

frontend/src/types/action-type.tsx

@@ -29,6 +29,9 @@ enum ActionType {
   // Logs a thought.
   THINK = "think",
 
+  // Calls an MCP tool
+  MCP_CALL_TOOL = "mcp_call_tool",
+
   // If you're absolutely certain that you've completed your task and have tested your work,
   // use the finish action to stop working.
   FINISH = "finish",

frontend/src/types/core/actions.ts

@@ -58,6 +58,16 @@ export interface FinishAction extends OpenHandsActionEvent<"finish"> {
   };
 }
 
+export interface MCPCallToolAction
+  extends OpenHandsActionEvent<"mcp_call_tool"> {
+  source: "agent";
+  args: {
+    thought: string;
+    tool_name: string;
+    kwargs: Record<string, unknown>;
+  };
+}
+
 export interface DelegateAction extends OpenHandsActionEvent<"delegate"> {
   source: "agent";
   timeout: number;
@@ -146,4 +156,5 @@ export type OpenHandsAction =
   | FileReadAction
   | FileEditAction
   | FileWriteAction
+  | MCPCallToolAction
   | RejectAction;

frontend/src/types/core/base.ts

@@ -12,6 +12,7 @@ export type OpenHandsEventType =
   | "reject"
   | "think"
   | "finish"
+  | "mcp_call_tool"
   | "error";
 
 interface OpenHandsBaseEvent {

frontend/src/types/core/observations.ts

@@ -109,6 +109,15 @@ export interface AgentThinkObservation
   };
 }
 
+export interface MCPCallToolObservation
+  extends OpenHandsObservationEvent<"mcp_call_tool"> {
+  source: "agent";
+  extras: {
+    tool_name: string;
+    kwargs?: Record<string, unknown>;
+  };
+}
+
 export type OpenHandsObservation =
   | AgentStateChangeObservation
   | AgentThinkObservation
@@ -120,4 +129,5 @@ export type OpenHandsObservation =
   | WriteObservation
   | ReadObservation
   | EditObservation
+  | MCPCallToolObservation
   | ErrorObservation;

frontend/src/types/observation-type.tsx

@@ -29,6 +29,9 @@ enum ObservationType {
   // A response to the agent's thought (usually a static message)
   THINK = "think",
 
+  // MCP tool call result
+  MCP_CALL_TOOL = "mcp_call_tool",
+
   // A no-op observation
   NULL = "null",
 }

openhands/agenthub/codeact_agent/function_calling.py

@@ -15,6 +15,7 @@
     FinishTool,
     IPythonTool,
     LLMBasedFileEditTool,
+    MCPCallTool,
     ThinkTool,
     WebReadTool,
     create_cmd_run_tool,
@@ -35,6 +36,7 @@
     FileEditAction,
     FileReadAction,
     IPythonRunCellAction,
+    MCPCallToolAction,
     MessageAction,
 )
 from openhands.events.event import FileEditSource, FileReadSource
@@ -191,6 +193,20 @@ def response_to_actions(response: ModelResponse) -> list[Action]:
                         f'Missing required argument "url" in tool call {tool_call.function.name}'
                     )
                 action = BrowseURLAction(url=arguments['url'])
+
+            # ================================================
+            # MCPCallTool (MCP)
+            # ================================================
+            elif tool_call.function.name == MCPCallTool['function']['name']:
+                if 'tool_name' not in arguments:
+                    raise FunctionCallValidationError(
+                        f'Missing required argument "tool_name" in tool call {tool_call.function.name}'
+                    )
+                action = MCPCallToolAction(
+                    tool_name=arguments['tool_name'],
+                    kwargs=arguments['kwargs'] if 'kwargs' in arguments else None,
+                )
+
             else:
                 raise FunctionCallNotExistsError(
                     f'Tool {tool_call.function.name} is not registered. (arguments: {arguments}). Please check the tool name and retry with an existing tool.'
@@ -245,6 +261,7 @@ def get_tools(
         create_cmd_run_tool(use_simplified_description=use_simplified_tool_desc),
         ThinkTool,
         FinishTool,
+        MCPCallTool,
     ]
     if codeact_enable_browsing:
         tools.append(WebReadTool)

openhands/agenthub/codeact_agent/tools/__init__.py

@@ -3,6 +3,7 @@
 from .finish import FinishTool
 from .ipython import IPythonTool
 from .llm_based_edit import LLMBasedFileEditTool
+from .mcp import MCPCallTool
 from .str_replace_editor import create_str_replace_editor_tool
 from .think import ThinkTool
 from .web_read import WebReadTool
@@ -16,4 +17,5 @@
     'create_str_replace_editor_tool',
     'WebReadTool',
     'ThinkTool',
+    'MCPCallTool',
 ]

openhands/agenthub/codeact_agent/tools/mcp.py

@@ -0,0 +1,33 @@
+from litellm import ChatCompletionToolParam, ChatCompletionToolParamFunctionChunk
+
+_MCP_CALL_DESCRIPTION = """Custom tool to call other tools exposed by MCP (Model Context Protocol) servers.
+
+For example, if you see a tool in the format: \
+"Tool(name='tool_name', description='Tool description', inputSchema={'type': 'object', 'properties':
+{'param1': {'type': 'string'}}})"
+
+You can call it by setting `tool_name` to `tool_name` and `kwargs` to \
+`{'param1': 'value'}`. \
+"""
+
+MCPCallTool = ChatCompletionToolParam(
+    type='function',
+    function=ChatCompletionToolParamFunctionChunk(
+        name='mcp_call_tool',
+        description=_MCP_CALL_DESCRIPTION,
+        parameters={
+            'type': 'object',
+            'properties': {
+                'tool_name': {
+                    'type': 'string',
+                    'description': 'The name of the tool to call.',
+                },
+                'kwargs': {
+                    'type': 'object',
+                    'description': 'The optional arguments dict to pass to the tool call.',
+                },
+            },
+            'required': ['tool_name'],
+        },
+    ),
+)

openhands/core/schema/action.py

@@ -83,3 +83,6 @@ class ActionType(str, Enum):
 
     CONDENSATION = 'condensation'
     """Condenses a list of events into a summary."""
+
+    MCP_CALL_TOOL = 'mcp_call_tool'
+    """Calls a MCP tool."""

openhands/core/schema/observation.py

@@ -49,3 +49,6 @@ class ObservationType(str, Enum):
 
     RECALL = 'recall'
     """Result of a recall operation. This can be the workspace context, a microagent, or other types of information."""
+
+    MCP_CALL_TOOL = 'mcp_call_tool'
+    """Result of a tool call to the MCP server."""

openhands/core/setup.py

@@ -162,6 +162,13 @@ def create_memory(
         )
         memory.load_user_workspace_microagents(microagents)
 
+        # get MCP tools from the runtime
+        mcp_tools_definition = runtime.get_mcp_tool_definitions(
+            memory.get_mcp_configs()
+        )
+        logger.info(f'All MCP tools: {mcp_tools_definition}')
+        memory.populate_mcp_tool_definitions(mcp_tools_definition)
+
         if selected_repository and repo_directory:
             memory.set_repository_info(selected_repository, repo_directory)
 

openhands/events/action/__init__.py

@@ -15,6 +15,7 @@
     FileReadAction,
     FileWriteAction,
 )
+from openhands.events.action.mcp import MCPCallToolAction
 from openhands.events.action.message import MessageAction
 
 __all__ = [
@@ -35,4 +36,5 @@
     'ActionConfirmationStatus',
     'AgentThinkAction',
     'RecallAction',
+    'MCPCallToolAction',
 ]

openhands/events/action/mcp.py

@@ -0,0 +1,19 @@
+from dataclasses import dataclass
+from typing import Any, ClassVar
+
+from openhands.core.schema import ActionType
+from openhands.events.action.action import Action
+
+
+@dataclass
+class MCPCallToolAction(Action):
+    runnable: ClassVar[bool] = True
+    thought: str = ''
+    tool_name: str = ''
+    kwargs: dict[str, Any] | None = None
+    action: str = ActionType.MCP_CALL_TOOL
+
+    @property
+    def message(self) -> str:
+        msg: str = f'Calling MCP tool `{self.tool_name}` with arguments: {self.kwargs}'
+        return msg

openhands/events/observation/__init__.py

@@ -21,6 +21,7 @@
     FileReadObservation,
     FileWriteObservation,
 )
+from openhands.events.observation.mcp import MCPCallToolObservation
 from openhands.events.observation.observation import Observation
 from openhands.events.observation.reject import UserRejectObservation
 from openhands.events.observation.success import SuccessObservation
@@ -44,4 +45,5 @@
     'AgentCondensationObservation',
     'RecallObservation',
     'RecallType',
+    'MCPCallToolObservation',
 ]

openhands/events/observation/mcp.py

@@ -0,0 +1,16 @@
+from dataclasses import dataclass
+from typing import Any
+
+from openhands.core.schema import ObservationType
+from openhands.events.observation.observation import Observation
+
+
+@dataclass
+class MCPCallToolObservation(Observation):
+    observation: str = ObservationType.MCP_CALL_TOOL
+    tool_name: str = ''
+    kwargs: dict[str, Any] | None = None
+
+    @property
+    def message(self) -> str:
+        return self.content

openhands/events/serialization/action.py

@@ -22,6 +22,7 @@
     FileReadAction,
     FileWriteAction,
 )
+from openhands.events.action.mcp import MCPCallToolAction
 from openhands.events.action.message import MessageAction
 
 actions = (
@@ -41,6 +42,7 @@
     ChangeAgentStateAction,
     MessageAction,
     CondensationAction,
+    MCPCallToolAction,
 )
 
 ACTION_TYPE_TO_CLASS = {action_class.action: action_class for action_class in actions}  # type: ignore[attr-defined]

openhands/events/serialization/observation.py

@@ -25,6 +25,7 @@
     FileReadObservation,
     FileWriteObservation,
 )
+from openhands.events.observation.mcp import MCPCallToolObservation
 from openhands.events.observation.observation import Observation
 from openhands.events.observation.reject import UserRejectObservation
 from openhands.events.observation.success import SuccessObservation
@@ -45,6 +46,7 @@
     AgentCondensationObservation,
     AgentThinkObservation,
     RecallObservation,
+    MCPCallToolObservation,
 )
 
 OBSERVATION_TYPE_TO_CLASS = {