MCP App Store
by KeenEthics
Driversnote icon

Driversnote MCP Server for Claude and ChatGPT

By Driversnote ApS

Overview

Use the Driversnote MCP to manage your trips, create and preview reports, submit reports to your team, and more!

How to connect

  1. Claude: Open the Driversnote listing in the Claude connectors directory and choose Connect. Claude listing Driversnote exposes 16 tools here.
  2. ChatGPT: Open the Driversnote listing in the ChatGPT app directory and choose Connect. ChatGPT listing Driversnote exposes 23 tools here.

Tools

classify_trips

ChatGPT
Update one or many trips in a single call: classify them as business or personal (set the trip type and workplace/organisation), change vehicle, set a tag or comment, and/or mark them as reviewed. This is the efficient way to classify trips — one request instead of one per trip. Example — classify last week's pending trips as business at a workplace, based on the user's calendar: { "ids": ["<trip ids from list_trips>"], "tripType": "business", "organisationId": "<workplace id from list_organisations>", "markAsReviewed": true } Example — mark trips as personal: { "ids": ["..."], "tripType": "private", "markAsReviewed": true } Notes: - All listed trips receive the SAME values. To give trips different classifications, group them and make one call per group. - Find trip IDs (and their current classification / reviewStatus) with list_trips first. - markAsReviewed: true sets reviewStatus to 'done', which removes the trips from the user's 'needs review' list. - Only one tag can be applied per bulk call (tagId, requires tagOrganisationId); attach multiple tags per trip at creation with create_trip instead.

create_location

ChatGPT
Resolve a place, given by its coordinates, to one of the user's saved locations — reusing an existing one when it's the same place, and creating it otherwise. Use it to turn trip endpoints from an import (Google Timeline, a GPS log, …) into saved locations before calling import_trips or create_trip. Example — a Google Timeline visit: { "lat": 55.6724, "lon": 12.5213, "googlePlaceId": "ChIJhwUoPJRTUkYR78kejeKf2zs" } Example — a calendar event location with a name: { "lat": 55.6724, "lon": 12.5213, "name": "Client HQ" } How it decides (checked on the server, in this order): 1. A saved location with the same Google place ID → reused (reuseReason: "place_id"). 2. A saved location within 100 m → reused ("nearby", reuseDistance in meters). 3. Otherwise the coordinates are looked up as an address. If that address is already saved → reused ("geocoded_place"). 4. Otherwise a new location is created with that address. Notes: - Call it ONCE per distinct place, not once per trip endpoint: group the import's endpoints first (same place ID, or within 100 m of each other). - name is optional. Only pass a name the source actually gives (a calendar location, a spreadsheet column, Timeline's HOME/WORK → "Home"/"Work"). Without a name, the app shows the address. - Address lookups are capped per user per day. Pass address, postalDistrictString and countryId when the source has them, to skip the lookup. - tripType gives the location a default classification for trips that start or end there.

create_report

ChatGPT
Create and save a mileage report for a period, vehicle and workplace. The report collects every matching trip, calculates distance and reimbursement at the workplace's rates, and appears in the user's report list with PDF and spreadsheet downloads. Use it when the user actually wants a report to keep, download or submit to their team — for 'what would it total?' questions use preview_report instead, which saves nothing. Recommended flow: preview_report with the same arguments first → check warnings → create_report → (team members) submit_report_to_team. Example: { "vehicleId": "<id>", "organisationId": "<id>", "startDate": "2026-03-01", "endDate": "2026-03-31" } Notes: - vehicleId/organisationId come from get_account_context; the inputs are identical to preview_report. - Check list_reports first if there is a risk the period is already reported: creating a second report over the same trips flags them as double-reported. - The returned pdfUrl is often null right after creation because the PDF is generated in the background — fetch it a little later with get_report_download_links. excelUrl is available immediately but requires a signed-in browser. - On the free plan the API refuses to create a report once the monthly trip limit is passed (preview_report's warnings.hasPassedMonthlyLimit). - Each call creates one report and costs one request of the 2,000/month quota.

create_tag

ChatGPT
Create a tag on a workplace, creating its tag category if needed. Reuses an existing tag with the same name in that category instead of creating a duplicate. Example — tag an import so the user can find it later: { "organisationId": "<workplace id>", "name": "Google Timeline import", "categoryName": "Imports" } Notes: - Pass categoryId (see list_tags) or categoryName. A category with that name is reused if it exists, and created if not. - On a team workplace only admins can create tags. If this fails with a permission error, use the existing tags from list_tags instead. - Returns the tag and whether it (reused) or its category (categoryCreated) already existed.

create_trip

ChatGPT
Create a single trip (mileage record) — ideal for importing past trips from a spreadsheet, calendar or another system. Call once per trip. Example — import one spreadsheet row: { "startAt": "2026-03-04T08:30:00+01:00", "stopAt": "2026-03-04T09:05:00+01:00", "distance": 23400, "startLocationPlaceholder": "Home, Copenhagen", "stopLocationPlaceholder": "Client HQ, Roskilde", "tripType": "business", "vehicleId": "<id from list_vehicles>", "organisationId": "<id from list_organisations>", "comment": "Imported from March spreadsheet" } Notes: - distance is an integer in the vehicle's local base unit: meters in km countries (23400 = 23.4 km), miles in mile countries. Check via get_account_context first. - vehicleId, organisationId and tripType are required — look up vehicleId/organisationId with get_account_context, list_vehicles and list_organisations before importing. - The reimbursement amount is calculated automatically from the workplace's rates. - Each of start/stop location is required: pass either the saved-location ID (startLocationId/stopLocationId, see list_locations) or a free-text placeholder (startLocationPlaceholder/stopLocationPlaceholder) — ID preferred when a saved location matches. - Each call costs one request of the 2,000/month quota, so confirm the mapping on 1-2 trips before importing a large batch.

create_vehicle

ChatGPT
Create a vehicle, e.g. a dedicated vehicle to import trips into ("Timeline", "Company car"). Example: { "name": "Timeline", "countryId": "DK", "vehicleTypeId": "<vehicleTypeId of an existing vehicle>" } Notes: - Take countryId and vehicleTypeId from an existing vehicle in get_account_context or list_vehicles. Vehicle types are per country, and the country decides the distance unit (km/mi) and currency of the vehicle's trips. - Check list_vehicles first, and reuse a vehicle with the same name instead of creating a duplicate. - odometerUsage controls how often the app asks for odometer readings: none (default), monthly, weekly or daily.

delete_trips

ChatGPT
Delete trips by ID, e.g. to undo an import that went wrong. Example — undo an import: { "ids": ["<createdTripIds from import_trips>"] } Notes: - ALWAYS confirm with the user before deleting, and say how many trips will be deleted. Never delete trips the user didn't ask to remove. - Deleted trips can be restored from the app's trash. - Find IDs with list_trips, or use the createdTripIds returned by import_trips.

get_account_context

ChatGPT
Bootstrap tool: fetch a compact overview of the account behind the API key in ONE request — vehicles (with country, distance unit km/mi, and currency), organisations (what the end user calls "workplaces"), teams the user belongs to (with their own role and name on each team), and the available reimbursement methods. Call this FIRST in a session. It gives you every ID you need for the other tools (vehicleId, organisationId, teamId, countryId, vehicleTypeId) plus the units and currency to interpret distances and amounts — while spending only one request of the monthly quota. Returns flat arrays: vehicles, organisations (call these "workplaces" when talking to the user), teams and reimbursementMethods. Notes: - The API key impersonates one user; each team's currentMemberRole (e.g. 'admin', 'manager' or 'member') is the closest the public API gets to 'who am I'. - If a list may exceed the built-in page sizes (25 vehicles/organisations, 10 teams), follow up with the dedicated list tools.

get_team_activity

ChatGPT
One-request overview of how a team is doing in a period — for managers/admins. Answers 'how is my team doing this month?', 'who hasn't submitted their mileage report?', 'who has stopped tracking?', 'how much mileage has the team reported for Q1?' — all from a single call. Example: { "startDate": "2026-03-01", "endDate": "2026-03-31" } (teamId optional; defaults to the user's default team, see get_account_context) Prefer passing the teamId from get_account_context: the API's fallback 'default team' can differ from the teams listed there. What comes back: - members: counts by role plus trackingSincePeriodStart / notTracking. The API only exposes each member's latest trip date, so 'tracking' means the member created a trip on or after startDate (their latest trip date is not before it); 'not tracking' means their latest trip is older than the period or they never tracked. - reports: team reports whose period overlaps the range, counted by state (pending = awaiting approval, approved, processed = paid out, rejected, unsubmitted = withdrawn draft), plus notSubmitted = drivers (role 'member') without a submitted report for the period. - submittedTotals: distance and amount across pending/approved/processed reports, grouped by currency + distance unit (rejected and unsubmitted reports are excluded). Distance is in distanceUnit (km or mi), amount in currency. - missingReports: the drivers who have not submitted for the period — name and email for follow-up. - memberActivity: one row per member with latest trip date, tracking flag, reports submitted in the period, their distance/amount and whether a report is missing. - complete is false when the team has more than 100 members or reports in the period; totals then cover only the first 100 — use summarize_team_reports (pages everything) and list_team_members for exact figures. Notes: - Requires the API key's owner to be a manager or admin of the team; members only see their own row. - Managers/admins are not expected to submit reports, so they never appear in missingReports. - For the money view alone (per-member payout totals with every report listed) use summarize_team_reports; for the raw report rows use list_team_reports.

guide

ChatGPT
Step-by-step guides for workflows that span several tools, with the Driversnote knowledge they need. Read the relevant guide BEFORE starting one of these workflows — e.g. before importing trips. Call without arguments to list the guides, or with a name to read one: { "name": "import-trips" } Guides: - import-trips: The general flow for importing past trips from any source — set-up, resolving locations, dry run, import, verify and undo. - import-google-timeline: Read Google Timeline (Maps location history) exports — the on-device Timeline.json and older Takeout files — and turn the car drives into trips. - import-calendar: Reconstruct trips from calendar events with a location, using distances from the user's earlier trips — never made-up ones. - import-spreadsheet: Map the columns of a spreadsheet, CSV or another mileage app's export (MileIQ, Everlance, TripLog, …) to trips. - classify-trips: Decide whether trips are business or private — using the calendar, saved locations and commuting rules — and update them in bulk.

import_trips

ChatGPT
Import many trips in one call — from Google Timeline, a calendar, a spreadsheet, another mileage app, or anything else the user has. Call the guide tool with "import-trips" (and the source-specific guide) before the first import. Always run it twice: 1. dryRun: true — validates every trip, checks for possible duplicates and returns a preview (summary, locations, possibleDuplicates, invalid) without writing anything. Show the user the preview and get their go-ahead. 2. The same trips with dryRun: false — creates them and returns createdTripIds (keep them: delete_trips with those IDs undoes the import). Example: { "dryRun": true, "trips": [ { "startAt": "2026-03-04T08:30:00+01:00", "stopAt": "2026-03-04T09:05:00+01:00", "distance": 23400, "startLocationId": "<id from create_location>", "stopLocationId": "<id from create_location>", "tripType": "business", "vehicleId": "<id>", "organisationId": "<id>", "tagIds": ["<id from create_tag>"], "comment": "Google Timeline (confidence 0.92)" } ] } Notes: - Each trip takes the same fields as create_trip. Resolve every endpoint to a saved location first (list_locations, or create_location for coordinates). Use a placeholder only for real address text — never coordinates. - A trip that overlaps in time with one of the user's existing trips (on any vehicle) is a possible duplicate. These are skipped and listed in possibleDuplicates, unless you pass importPossibleDuplicates: true after the user has confirmed they want them. - At most 100 trips per call; split bigger imports into several calls. - Imported trips are marked as imported, are never classified automatically, and are left for the user to review (reviewStatus 'pending') unless you pass markAsReviewed: true. - Driversnote accepts trips from the last 5 years (and up to 1 year ahead). Older trips are listed in invalid. - distance is in the vehicle's base unit: meters in km countries, miles in mile countries. - Suggest tagging the import (e.g. "Google Timeline import", via create_tag) so the user can find and filter it later.

list_locations

ChatGPT
List the user's saved locations (named places like 'Home', 'Office', 'Client HQ' used on trips), with address and coordinates. Use this to find startLocationId/stopLocationId values for create_trip and classify_trips — linking trips to saved locations is better than free-text placeholders. Filter with searchPhrase to match a name or address substring.

list_organisations

ChatGPT
List the user's organisations — what the end user and the Driversnote app call "workplaces". Returns name, address, whether reimbursement is calculated, and the reimbursement method in use. An organisation/workplace is what a business trip is driven FOR — assigning one to a trip is what classifies it as business for that employer/client. Use this to find the organisationId needed by create_trip, classify_trips, list_trips and create-report flows. When talking to the end user, call these "workplaces", not "organisations" — that's the term they know. Set isArchived: false to only list active organisations (default returns all). Each row flattens the reimbursement method into reimbursementMethodId/reimbursementMethodName/reimbursementMethodDescription.

list_reports

ChatGPT
List the user's own mileage reports (documents that collect trips for a period, used for reimbursement or tax documentation), newest data included: period, total distance and amount, currency, state, workplace, vehicle, and download URLs for PDF/Excel. Use this to answer 'what did my March report total?', to find a report's PDF link, or to check which periods already have a report before creating new ones. Notes: - These are the user's personal reports. For reports submitted by team members to a team the user manages, use list_team_reports instead. - state is one of: not_created, created, unsubmitted, pending, approved, rejected, processed (states beyond 'created' reflect team submission). - totalDistance is in the report's distanceUnit; totalAmount is in currency. - pagination.total is the number of reports across all pages; paginate with first/after using pagination.nextCursor.

list_tags

ChatGPT
List the tags (and tag categories) of a workplace. Tags label trips, e.g. a project, a client or "Google Timeline import", and can be filtered on and shown in reports. Use this to find tagIds for create_trip and import_trips, and tagId for classify_trips. Tags belong to a workplace, so a trip can only carry tags from its own workplace. Example: { "organisationId": "<workplace id from get_account_context>" } Returns categories and a flat tags list where each tag carries its categoryId/categoryName. Create missing tags with create_tag. complete is false when the workplace has more categories or tags than could be read. A workplace that has never changed its tags lists Driversnote's default tags. They are copied to the workplace the first time one is used, which gives them new IDs: the old IDs keep working, but call list_tags again rather than reusing IDs from an earlier conversation.

list_team_members

ChatGPT
List the members of a team the user belongs to — name, email, role (member, manager or admin), license status and the date of each member's latest tracked trip. This is the tool for 'who is on my team?', for finding a member's memberId to filter list_team_reports / summarize_team_reports, or a manager's ID to use as reviewerId in submit_report_to_team. Notes: - teamId is optional and defaults to the user's default (first) team; find team IDs with get_account_context. The team itself is returned as team. Prefer passing the teamId from get_account_context: the API's fallback 'default team' can differ from the teams listed there. - Filter by role: 'member' (drivers who submit reports), 'manager' or 'admin' (who approve them). - Email, license and activity fields (userEmail, userHasLicense, userHasTeamLicense, userLatestTripCreatedAt, userHasPassedMonthlyLimit) are only returned when the API key's owner is a manager/admin of the team (or for the owner's own row); otherwise they are null. - userLatestTripCreatedAt is the best available activity signal: null means the member has never tracked a trip. - For a period-based activity overview (who tracked, who has not submitted a report) use get_team_activity instead. - pagination.total is the member count across all pages; paginate with first/after using pagination.nextCursor.

list_team_reports

ChatGPT
List mileage reports submitted to a team the user manages or administers — with submitting member (name/email), approval state, report period, total distance and reimbursement amount. This is the tool for team mileage analysis and approval workflows. Use it to answer 'how much mileage did my team submit in Q1?', 'which reports are still pending approval?', or 'what does the company owe each driver for March?' (group by teamMember, sum report.totalAmount). Notes: - The API key impersonates its owner: team reports are only visible if that user is a manager or admin of the team. A permissions error usually means the key's owner lacks that role. - teamId is optional and defaults to the user's first team; find team IDs with get_account_context. - Filter by state: unsubmitted, pending, approved, rejected, processed. - startDate/endDate filter on the report period. - memberEmail is only returned to team managers/admins. - Each row flattens the submitting member to memberId/memberName/memberEmail and keeps the report itself under report. - pagination.total is the number of team reports across all pages; paginate with first/after using pagination.nextCursor.

list_trips

ChatGPT
List the user's trips (mileage records) in a date range. Each trip answers: WHEN (startAt/stopAt), WHERE (startLocationName/startLocationAddress and stopLocationName/stopLocationAddress for saved locations, otherwise the free-text startLocationPlaceholder/stopLocationPlaceholder), HOW FAR (distance), WHAT IT IS WORTH (amount) and WHICH KIND (tripType: business/private/…). These are the attributes that matter for nearly every question — start there. Use this to answer 'how many business km did I drive in Q2?' (sum distance), 'where did I drive on the 12th?', to find unclassified trips (reviewStatus 'pending' = still needs review), or to look up trip IDs before classifying them with classify_trips. Notes: - startDate/endDate are required ISO dates, e.g. { "startDate": "2026-04-01", "endDate": "2026-06-30" }. - fields controls how much each row carries. 'essential' (default) = id, startAt, stopAt, the location fields, distance, amount, tripType, reviewStatus, comment — enough for almost everything and cheapest in tokens. 'full' adds odometer readings, routeOrigin, vehicle, workplace (organisation), tags and the trip type's display name — ask for it only when a question needs those. - distance is an integer in the vehicle's local base unit: meters in km countries (23400 = 23.4 km), miles in mile countries (check distanceUnit via get_account_context). - amount is the reimbursement value in the workplace's currency. - reviewStatus is one of: not_ready, pending, done. - Filter by tripType (business, private, medicalOrMoving or charitable), vehicleId, or organisationId (a workplace). - summary.count is the total number of matching trips across all pages. summary.distanceSum/amountSum cover only the trips returned so far; when summary.complete is true this page holds every match and the sums are final, otherwise keep paging with pagination.nextCursor and add up the sums yourself. - Paginate with first/after using pagination.nextCursor; keep pages large (e.g. first: 100) to conserve the 2,000 requests/month API quota.

list_vehicles

ChatGPT
List the user's vehicles with name, license plate, vehicle type (e.g. car, motorcycle, bicycle), odometer usage setting, and the vehicle's country (which determines distance unit — km or mi — and currency). Use this to find the vehicleId needed by create_trip, classify_trips and list_trips, and to check whether trip distances are in meters (km countries) or miles. Set isArchived: false to only see active vehicles (default returns all). Each row flattens the vehicle type and country into vehicleTypeId/vehicleTypeName, countryId/countryName, distanceUnit and currency.

preview_report

ChatGPT
Calculate what a mileage report WOULD contain for a period, vehicle and workplace — total distance and reimbursement amount, one summary line per rate/trip type, and data-quality warnings — without creating a report. Nothing is saved, so this is the tool for on-the-fly questions like 'what would I be reimbursed for March?', 'how much business mileage did I do in Q2 at my workplace's rate?' or for checking a period before calling create_report — it avoids cluttering the user's report list with throwaway reports. Example: { "vehicleId": "<id>", "organisationId": "<id>", "startDate": "2026-03-01", "endDate": "2026-03-31" } Get vehicleId/organisationId from get_account_context. Pass the exact same arguments to create_report when the user wants the report saved. Reading the result: - totals.distance (in distanceUnit: km or mi) and totals.amount (in currency) are the report's grand totals. - lines break the total down. type is 'rate' for a reimbursed rate (description = the rate's name, e.g. 'Business, first 20,000 km'), a trip type key such as 'business'/'private' when the workplace does not reimburse, 'none' for trips without a rate, or 'gap' for an odometer gap. percentage is the line's share of the distance. - warnings flag problems the user should fix before creating the report: odometer inconsistencies/missing readings, trips already included in another report (hasDoubleReportedTrips), archived trips, a provisional (not yet confirmed) rate, or the free-plan monthly trip limit (hasPassedMonthlyLimit — create_report will be refused). - hasReimbursementMethod is false when the workplace has no reimbursement method, in which case amounts are 0. - For the individual trips behind these numbers call list_trips with the same startDate/endDate/vehicleId/organisationId. - Costs one request of the 2,000/month quota.

submit_report_to_team

ChatGPT
Submit one of the user's mileage reports to their team for approval and reimbursement. Only relevant when the user is a MEMBER of a team (see teams[].currentMemberRole in get_account_context); the team's managers are notified by email and the report shows up in their approval queue as 'pending'. Example: { "reportId": "<id from create_report or list_reports>", "note": "March mileage, incl. the Aarhus client visits" } Notes: - teamId is optional: when the user belongs to exactly one team it is resolved automatically (one extra request); if they belong to several, pass the teamId from get_account_context. - The report must belong to the user (list_reports). Reports in state 'created' or 'unsubmitted' can be submitted; a report that was rejected can be re-submitted after the user fixes it. - note is shown to the reviewer alongside the report. reviewerId (a team member ID, see list_team_members) requests a specific manager as reviewer; omit to let the team's default routing apply. - Confirm with the user before submitting — this notifies other people and cannot be undone through this API. - The result's teamReport.state will be 'pending'.

summarize_team_reports

ChatGPT
Aggregate a team's mileage reports for a period into totals per member and per state — the payroll/reimbursement view. Answers 'what do we owe each driver for March?', 'how much mileage has the team claimed this quarter?', 'how many reports are still pending, and for what amount?'. Unlike list_team_reports, this pages through EVERY matching report before summing, so the totals are complete (see complete). Example — approved but not yet paid out, ready for payroll: { "startDate": "2026-03-01", "endDate": "2026-03-31", "state": "approved" } Reading the result: - totals: report count, distance and amount across all matching reports. Distance is in distanceUnit (km or mi), amount in currency. If the team mixes currencies or units, currency/distanceUnit are null and byCurrency holds one entry per combination — never add those together. - byState: count, distance and amount per state (pending = awaiting approval, approved, processed = paid out, rejected, unsubmitted = withdrawn draft). - byMember: one row per driver with their report count, distance, amount, a per-state count and the individual reports (id, title, period, state, amount, pdfUrl) so you can list or download them. Sorted by amount, highest first. - Rejected and unsubmitted reports are included in counts but NOT in distance/amount totals, so totals reflect what was actually claimed. Notes: - teamId defaults to the user's default team (see get_account_context). Requires the API key's owner to be a manager/admin of the team. Prefer passing the teamId from get_account_context: the API's fallback 'default team' can differ from the teams listed there. - startDate/endDate match reports whose period overlaps the range; omit both to summarize all of the team's reports. - Costs one request per 100 reports, at most 10. For a period-based engagement overview (who is tracking, who is missing a report) use get_team_activity.

classify_trips

Claude

create_report

Claude

create_trip

Claude

get_account_context

Claude

get_team_activity

Claude

list_locations

Claude

list_organisations

Claude

list_reports

Claude

list_team_members

Claude

list_team_reports

Claude

list_trips

Claude

list_vehicles

Claude

preview_report

Claude

submit_report_to_team

Claude

summarize_team_reports

Claude

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/driversnote"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/driversnote)