Felt MCP
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Felt MCP list my maps"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Felt MCP
An MCP server for interacting with Felt maps programmatically.
Prerequisites
Python 3.14+
Related MCP server: map-mcp
Configuration
Set your Felt API token as an environment variable:
export FELT_API_TOKEN=felt_pat_...Running locally
uv sync
uv run python -m felt_mcp.mainThe server uses stdio transport by default, which is the standard for MCP servers launched as subprocesses by an MCP client.
Running with Docker
docker build -t felt-mcp .
docker run -e FELT_API_TOKEN felt-mcpMCP client configuration
Add this to your MCP client config (e.g. Claude Desktop claude_desktop_config.json):
{
"mcpServers": {
"felt": {
"command": "uv",
"args": ["run", "--directory", "/path/to/felt-mcp", "python", "-m", "felt_mcp.main"],
"env": {
"FELT_API_TOKEN": "felt_pat_..."
}
}
}
}Or with Docker:
{
"mcpServers": {
"felt": {
"command": "docker",
"args": ["run", "-i", "--rm", "-e", "FELT_API_TOKEN", "felt-mcp"],
"env": {
"FELT_API_TOKEN": "felt_pat_..."
}
}
}
}Available tools
list_maps
Lists all maps accessible to the authenticated Felt user. Returns maps from all projects with their id, title, URL, project, and access level.
get_map_details
Get details for a specific Felt map, including its layers, elements, and settings.
Parameter | Type | Required | Description |
| string | Yes | The ID of the map to retrieve |
get_map_layers
List all layers on a specific Felt map.
Parameter | Type | Required | Description |
| string | Yes | The ID of the map whose layers to list |
create_new_map
Create a new Felt map with an optional title and initial viewport.
Parameter | Type | Required | Description |
| string | No | Title for the new map |
| float | No | Initial latitude |
| float | No | Initial longitude |
| int | No | Initial zoom level |
Development
uv sync # install all dependencies including dev
just lint # format, lint, and type check
just test # run test suite
just test-failed # re-run previously failed testsLicense
See LICENSE.md.
Available Tools
4 toolscreate_new_mapA
Create a new Felt map.
Args: title: Optional title for the new map. lat: Optional initial latitude. lon: Optional initial longitude. zoom: Optional initial zoom level.
Inputs: lat, lon, zoom, title. Returns: object
get_map_detailsA
Get details for a specific Felt map, including its layers and elements.
Args: map_id: The ID of the map to retrieve.
Inputs: map_id (required). Returns: object
get_map_layersA
List all layers on a specific Felt map.
Args: map_id: The ID of the map whose layers to list.
Inputs: map_id (required). Returns: result
list_mapsA
List all maps accessible to the authenticated Felt user.
Returns a list of maps with their id, title, URL, project, and access level.
Inputs: no named parameters. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 4 tools
get_map_details and get_map_layers overlap since details already includes layer information, but one is a general map overview and the other specifically returns only layers. list_maps and create_new_map are clearly distinct.
All tool names follow a predictable verb_noun snake_case pattern: get_map_details, get_map_layers, create_new_map, list_maps. The small variations ('new' and plural 'maps') do not create confusion.
Four tools is a compact, well-scoped set for a map-oriented MCP server. Each tool covers a reasonable core action: list, get details, get layers, and create.
The set covers list/get/create for maps and layer listing, but lacks update/delete for maps and any layer manipulation such as create, update, or delete. This leaves the map lifecycle incomplete, though common read-and-create workflows are supported.
Maintenance
Related MCP Connectors
Hosted MCP for live place search, geocoding & turn-by-turn routes. No Maps API key to manage.
MCP server for interacting with the Supabase platform
Hosted MCP server for AI-driven data ops. Create apps, manage schemas, and CRUD structured data.
Nifty's MCP server — exposes tasks, projects, messages, and files as tools for AI agents.
Related MCP Servers
- MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to perceive and control a live MapLibre GL map over MCP, allowing querying rendered features, reading popups, navigating, and toggling layers.1MIT
- FlicenseNot gradedqualityDmaintenanceComprehensive MCP server for Google Maps APIs, enabling geocoding, place search and details, distance matrix, elevation, and directions through natural language.6-
- FlicenseNot gradedqualityDmaintenanceMCP server that drives Azgaar's Fantasy Map Generator in a real browser via Playwright, enabling map generation, inspection, and exporting screenshots/SVG/JSON through natural language.1-