rpcclient-mcp
MCP server for iOS automation via rpcclient, enabling AI agents to connect to iOS RPC servers and perform actions like launching apps, clicking, typing, and gesture control.
README
rpcclient-mcp
MCP server for rpcclient targets iOS over stdio transport.
Prerequisites
- Python 3.10+
- RPC server reachable
Getting Started
Install
python3 -m venv .venv
source .venv/bin/activate
pip install rpcclient-mcp
which rpcclient-mcp
rpcclient==6.11.1 is a package dependency, so it installs automatically with this MCP.
If you want to force local rpcclient source instead of PyPI:
export RPCCLIENT_PYTHONPATH=/absolute/path/to/rpc-project/src/rpcclient
Configure your AI agent
Claude Code
claude mcp add --transport stdio rpcclient-mcp -- /absolute/path/to/venv/bin/rpcclient-mcp
Codex CLI
codex mcp add rpcclient-mcp -- /absolute/path/to/venv/bin/rpcclient-mcp
Gemini CLI
gemini mcp add --transport stdio rpcclient-mcp /absolute/path/to/venv/bin/rpcclient-mcp
Codex (manual config)
Add to ~/.codex/config.toml (or your active Codex config):
[mcp_servers.rpcclient-mcp]
command = "/absolute/path/to/venv/bin/rpcclient-mcp"
[mcp_servers.rpcclient-mcp.env]
RPCCLIENT_PYTHONPATH = "/absolute/path/to/rpc-project/src/rpcclient" # optional
Restart your client after config changes.
Usage
ping()
rpc_connect(host="127.0.0.1", port=5910)
rpc_get_screen_context(session_id="...")
rpc_disconnect(session_id="...")
Tools
Session
# health
ping()
# connection lifecycle
rpc_connect(host="127.0.0.1", port=5910, timeout=3.0)
rpc_disconnect(session_id)
rpc_reconnect(session_id)
# session inspection
rpc_list_sessions()
rpc_session_info(session_id)
Use these to establish, inspect, and recover sessions.
rpc_reconnect is the transport recovery path and invalidates old handles.
Generic RPC + Handles
# discovery
rpc_capabilities(session_id, include_private=False)
rpc_call(session_id, path, args=None, kwargs=None, store_result=True)
# handle lifecycle
rpc_list_handles(session_id)
rpc_release_handle(session_id, handle_id)
rpc_release_all_handles(session_id)
# handle access
rpc_handle_call(session_id, handle_id, method, args=None, kwargs=None, store_result=True)
rpc_handle_get(session_id, handle_id, attr, store_result=True)
Use rpc_call for dot-path execution on the root client object.
When store_result=True, non-JSON return values are materialized as handles.
App + UI
# app discovery + launch
rpc_list_apps(session_id, query="", limit=30)
rpc_launch_app(session_id, bundle_id=None, app_query=None, kill_existing=True, unlock_device=True, timeout=3.0, wait_foreground_timeout=5.0)
# context + overlay handling
rpc_get_screen_context(session_id, ensure_accessibility=True, wait_primary_bundle=None, wait_timeout=5.0, max_items=300, dedupe=True)
rpc_dismiss_overlays(session_id, timeout=4.0, max_rounds=6)
# interaction: click + type
rpc_click_label(session_id, label, exact=True, index=0, timeout=5.0, auto_scroll=True, draw_frame=False, direction="next", displayed_only=False, dismiss_overlays=True, overlay_timeout=3.0)
rpc_click_by_context(session_id, intent, min_score=0.52, timeout=5.0, auto_scroll=True, draw_frame=False, direction="next", dismiss_overlays=True, overlay_timeout=3.0)
rpc_type_text(session_id, text, target_label=None, intent=None, min_intent_score=0.35, anchor_label=None, anchor_position="before", anchor_offset=1, clear_first=False, timeout=5.0, direction="next", displayed_only=False, dismiss_overlays=True, overlay_timeout=3.0)
# gestures + focused fallback
rpc_scroll(session_id, direction="down", count=1, pause_seconds=0.25, from_x=None, from_y=None, to_x=None, to_y=None)
rpc_tap(session_id, x, y, repeat=1, pause_seconds=0.2)
rpc_type_focused_text(session_id, text, dismiss_overlays=False, overlay_timeout=2.0)
This is the primary automation surface: launch, read context, click, type, and gesture.
Use rpc_tap + rpc_type_focused_text when web fields are visible but not AX-discoverable.
Current Limitation
Safari overlay password fields are not always AX-exposed on this target.
Manual password entry may still be required in some runs.
License
MIT. See LICENSE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.