Virtual location (Google Street View,etc.)

Virtual location (Google Street View,etc.)

An MCP server that enables a virtual travel avatar on Google Maps with image outputs.

23
Stars
12
Forks
20
Releases

Overview

An MCP server that creates an environment for an avatar to virtually travel on Google Maps. It supports MCP clients (e.g., Claude Desktop) to issue instructions and receive journey progress with photos. The server can run as stdio or Streamable-HTTP (Smithery.ai config compatible) and supports multiple concurrent sessions, with per-session data stored via Turso libsql or local sqlite. It integrates Google Maps APIs (Street View, Places, Time Zone, Directions) to provide real-world location context and nearby facilities, enhancing travel realism. A comprehensive set of MCP tools allows querying and controlling the traveler’s state: location, destination, progress, and journey start/stop; avatar appearance is configurable via prompts, with a default anime style and nano-banana-capable image generation. Image backends include Gemini nano-banana, PixAI, Stability.ai, and optional ComfyUI, enabling flexible image synthesis workflows. Bluesky SNS integration enables reading and posting to Bluesky feeds, including likes and replies. The project supports practice mode, per-session base64 configuration, per-user databases, and optional remote usage with Smithery, MseeP verification, and librechat. It emphasizes configurable output controls (MT_NO_IMAGE, MT_NO_AVATAR) and adjustable image width for cost/quality balance.

Details

Owner
mfukushim
Language
TypeScript
License
MIT License
Updated
2025-12-07

Features

Traveler management and journey control

Expose functions to query and set the traveler’s current location, destination, progress, and to start/stop the journey (e.g., get_traveler_location, set_traveler_location, set_traveler_destination_address, start_traveler_journey, stop_traveler_journey, reach_a_percentage_of_destination).

Avatar prompt and appearance control

Manage avatar image prompts with set_avatar_prompt; supports a default anime style and nano-banana-based generation to optimize composite images.

MCP server tools and state queries

Provide access to traveler state and environment through functions like get_traveler_view_info, get_traveler_info, and get_setting for dynamic responses.

Bluesky SNS integration

Read Bluesky feeds/mentions and post/reply/like via post_sns_writer, reply_sns_writer, and related tools, with tag-based routing.

Practice mode and deployment options

Supports practice mode (no required API keys) and multiple deployment modes (stdio and streamable-http), with Smithery compatibility.

Image generation backends

Supports Gemini nano-banana for fast composites; optional PixAI, Stability.ai; and optional ComfyUI as a local image-generation server.

Config and environment flexibility

Extensive environment variables, per-session base64 config, Turso libsql DB support, and options to control image output and avatar overlay.

Map data integration

Uses Google Map APIs (Street View, Places, Time Zone, Directions) to place the avatar and fetch nearby facilities for realistic travel context.

Audience

MCP clientsInteract with travel avatar via LLM prompts and receive image updates.
DevelopersConfigure, deploy, and customize the MCP server settings, environment variables, and per-session configs.
LLM usersOperate the travel bot using natural language prompts to guide journeys.

Tags

MCPGoogle MapsStreet ViewBlueskyimage-generationavatarvirtual-travelnano-bananaPixAIStability.aiComfyUISmitheryClaude DesktopLibreChatTursorembg