MCP SERVER · OFFICIAL PROJECTAsk about your own recorded life over a time range — workouts, sleep cycles, where you were, and any metric or annotation you have been collecting — and get it back as raw samples or as a resampled series.
LABSCO SUMMARY
What this server can tell you is exactly what you have been recording, and the two catalog calls are how you find that out before writing a query. The value is in the shaping done before the data is handed over — sleep comes back reconstructed into cycles rather than as raw stages, and a metric can arrive resampled so a model is never asked to average samples itself. It reads and does not write: there is no call here for recording an annotation or correcting a sample.
A read layer over Context by Fulcra, the personal data store that collects your health, location and self-recorded data. It exposes catalogs of what you have been recording plus time-bounded reads of workouts, sleep, location and metrics. A public instance is hosted at https://mcp.fulcradynamics.com/mcp; this package is for running it yourself.
A Context by Fulcra account holding the data, authorised over OAuth. Point an OAuth-capable client at the public instance and install nothing, or run it locally with uvx — the local server handles the OAuth callback itself and keeps the exchanged tokens rather than passing them to the client.
One command — uvx fulcra-context-mcp@latest
USE CASES
Pull the night as cycles rather than as raw stages, then line it up against the workouts and the moments recorded either side of it.
Read the catalog of metrics and the catalog of annotations first — they name what this account has actually been collecting.
Ask for the series at a fixed sample rate instead of pulling every raw sample and reducing them in the model.
Pull the underlying samples for that window — anything whose own range overlaps it is included, so a long sample straddling the boundary still shows up.
Resolve the location at that exact moment, then widen out to the track across the hours around it.
Read your own time zone and calendar IDs first, so the results are interpreted in local time rather than in whatever the client assumes.
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
annotations_catalogGet the list of all annotations the user has defined. This does not get the
actual values the user has recorded; for that, use the get_annotations tool.
Use this tool to get the IDs and types to pass to get_annotations.
get_annotationsRetrieve an array of all moment annotations the user recorded during a period of time. Each item contains the value (except for moment annotations) and the metadata (name, original spec, etc.) describing the annotation.
Args: ann_type: annotation type (moment, duration, boolean, numeric, scale, etc.) start_time: The starting time of the period. Must include tz (ISO8601). end_time: the ending time of the period. Must include tz (ISO8601).
get_location_at_timeGets the user's location at the given time.
If no sample is available for the exact time, searches for the closest one up to window_size seconds back.
Result timestamps will include time zones. Always translate timestamps to the user's local time zone when this is known.
Args: time: The point in time to get the user's location for. Must include tz (ISO8601). window_size: Optional. The size (in seconds) to look back (and optionally forward) for samples. Defaults to 14400. include_after: Optional. When true, a sample that occurs after the requested time may be returned if it is the closest one. Defaults to False. Returns: A JSON string representing the location data.
get_location_time_seriesRetrieve a time series of locations that the user was at. Result timestamps will include time zones. Always translate timestamps to the user's local tz when this is known.
Args: start_time: The start of the time range (inclusive), as an ISO 8601 string or datetime object. end_time: The end of the range (exclusive), as an ISO 8601 string or datetime object. change_meters: Optional. When specified, subsequent samples that are fewer than this many meters away will not be included. sample_rate: Optional. The length (in seconds) of each sample. Default is 900. reverse_geocode: Optional. When true, Fulcra will attempt to reverse geocode the locations and include the details in the results. Default is False. Returns: A JSON string representing a list of location data points.
get_metric_samplesRetrieve the raw samples related to a given metric for the user during a specified period.
In cases where samples cover ranges and not points in time, a sample will be returned if any part of its range intersects with the requested range. For example, if start_time is 14:00 and end_time is 15:00, a sample covering 13:30-14:30 will be included. Result timestamps will include time zones. Always translate timestamps to the user's local time zone when this is known.
Args: metric_name: The name of the metric to retrieve samples for. Use get_metrics_catalog to find available metrics. start_time: The start of the time range (inclusive), as an ISO 8601 string or datetime object. end_time: The end of the time range (exclusive), as an ISO 8601 string or datetime object. Returns: A JSON string representing a list of raw samples for the metric.
get_metric_time_seriesGet user's time-series data for a single Fulcra metric.
Covers the time starting at start_time (inclusive) until end_time (exclusive). Result timestamps will include tz. Always translate timestamps to the user's local tz when this is known.
Args: metric_name: The name of the time-series metric to retrieve. Use get_metrics_catalog to find available metrics. start_time: The starting time period (inclusive). Must include tz (ISO8601). end_time: The ending time (exclusive). Must include tz (ISO8601). sample_rate: Optional. The number of seconds per sample. Default is 60. Can be smaller than 1. replace_nulls: Optional. When true, replace all NA with 0. Default is False. calculations: Optional. A list of additional calculations to perform for each time slice. Not supported on cumulative metrics. Options: "max", "min", "delta", "mean", "uniques", "allpoints", "rollingmean". Returns: A JSON string representing a list of data points for the metric. For time ranges where data is missing, the values will be NA unless replace_nulls is true.
get_metrics_catalogGet the catalog of available metrics that can be used in time-series API calls
(metric_time_series and metric_samples).
get_sleep_cyclesReturn sleep cycles summarized from sleep stages.
Processes raw sleep data samples into sleep cycles by finding gaps in the sleep sample data within a specified time interval. Result timestamps will include time zones. Always translate timestamps to the user's local time zone when this is known.
Args: start_time: The starting timestamp (inclusive), as an ISO 8601 string or datetime object. end_time: The ending timestamp (exclusive), as an ISO 8601 string or datetime object. cycle_gap: Optional. Minimum time interval separating distinct cycles (e.g., "PT2H" for 2 hours).
get_user_infoReturn general info about the Context by Fulcra user.
Returns user references such as time zone, calendar ids, and other metadata.
get_workoutsGet details about the workouts that the user has done during a period of time. Result timestamps will include time zones. Always translate timestamps to the user's local time zone when this is known.
Args: start_time: The starting time of the period. Must include tz (ISO8601). end_time: the ending time of the period. Must include tz (ISO8601).
BEFORE YOU INSTALL
10 tools when we connected.
uvx on PATH
CONTEXT COST
Rated by the size of the 10 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Add the block to your client, or run the command directly.
{
"mcpServers": {
"fulcra-context-mcp": {
"command": "uvx",
"args": [
"fulcra-context-mcp@latest"
]
}
}
}uvx fulcra-context-mcp@latestMCP servers
a1c-ai-agent
Monasterolo21
fliptheweb
khaoss85