Tempo MCP
Ingests revenue events from Stripe to compute business metrics like Revenue per Recovery Hour.
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., "@Tempo MCP Log checkin: energy 3, stress 8"
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.
Tempo MCP
The founder-vitals connector for solo operators. Tempo is the first MCP server that joins human capacity (sleep, mood, stress, energy) to business load (meetings, revenue, deal flow) and exposes them to Claude as one signal.
For a solopreneur, the founder is the single point of failure. Every other tool treats the business and the human as separate things. Tempo fuses them, so Claude can make business decisions — decline meetings, reprice services, prescribe rest days — based on the founder's real-time state.
What makes this novel
Tempo invents a single KPI that no other tool can compute on its own:
Revenue per Recovery Hour (RpRH) — revenue generated divided by hours of true recovery. Falling RpRH means you're working harder for less, and it's the earliest quantitative signal of founder burnout.
Related MCP server: cadence
Install
pip install tempo-mcpRequires Python 3.10+ and works with both mcp 1.x and 2.x.
Register with Claude Desktop / Cowork / Claude Code
~/.claude/settings.json (or equivalent):
{
"mcpServers": {
"tempo": {
"command": "python",
"args": ["-m", "tempo_mcp"]
}
}
}Tools exposed over MCP
Tool | Purpose |
| Log mood / energy / stress / sleep |
| Record revenue (from Stripe connector or manual) |
| Record a calendar block (internal, sales, deep work) |
| Compute the composite Tempo score + RpRH |
| Prescriptive business moves based on current state |
| Log (or, on Pro, execute) an action |
| Full time-series history over N days — for trend analysis |
| Narrative week-over-week summary Claude can read aloud |
| Show the active plan and what it unlocks |
| Activate a Pro / Team Solo key |
| Export the whole local dataset as JSON or CSV (Pro) |
Plans
Tempo is free forever for the core loop. Paid tiers exist because the expensive parts — long history, connector ingest, executing actions — are the parts that need maintaining.
Free | Pro — $39/mo | Team Solo — $79/mo | |
Daily check-ins, Tempo score, RpRH | ✅ | ✅ | ✅ |
Manual revenue / calendar logging | ✅ | ✅ | ✅ |
History window | 7 days | unlimited | unlimited |
Week-over-week digest | — | ✅ | ✅ |
Connector ingest (Stripe, calendar, wearable) | — | ✅ | ✅ |
Executed actions | log only | 3 / day | unlimited |
Data export (JSON / CSV) | — | ✅ | ✅ |
Seats | 1 | 1 | 5 |
Subscribe at temposolo.app, then paste the key from your welcome email:
activate_license with key tempo_v1....Or set it yourself — TEMPO_LICENSE_KEY in the environment, or one line
in ~/.tempo/license.
Licence keys are Ed25519-signed and verified offline against a public key baked into the package. There is no licence server, no activation call, and no telemetry: Tempo works on a plane and cannot tell us when you use it. An expired or missing key quietly drops you to the free tier rather than locking you out of your own data.
Example conversation with Claude
You: "I slept 5 hours, energy's at 3/10, stress 8. Log it and tell me what to do about tomorrow."
Claude (using Tempo): logs checkin → pulls score → RpRH down 22% → 4 external calls booked tomorrow → recommends auto-decline of 2 discovery calls and a rest day.
Data stays local
Tempo stores everything in ~/.tempo/tempo.sqlite3. Nothing leaves your machine unless you enable the Pro cloud sync.
Develop
git clone https://github.com/tsutera09/tempo-mcp
cd tempo-mcp
pip install -e '.[dev]'
pytest
python -m tempo_mcp # runs the MCP server on stdioPublishing
See PUBLISH.md at the repo root for the ordered release checklist (PyPI, npm mirror, MCP registry).
License
MIT
Available Tools
11 toolsactivate_licenseA
Activate a Tempo Pro or Team Solo licence key.
Paste the key from the welcome email. It is verified offline against Tempo's signing key and saved to ~/.tempo/license — nothing is sent anywhere.
Inputs: key (required).
apply_actionA
Record — or on Pro, execute — an action Tempo recommended.
Args: kind: auto_decline, reprice, rest_day, block_deep_work, ... payload: the action's parameters (who, when, how much) outcome: applied | dismissed | pending execute: Pro only. Hand the action to the connected executor (Cal.com / Stripe / Gmail) instead of only writing it down. Pro allows 3 executions per day; Team Solo is unlimited.
Logging is always free — the decision record is the point of the free tier, and it is what makes the weekly retrospective honest.
Inputs: kind (required), payload (required), execute, outcome.
export_dataA
Export the full local Tempo dataset. Pro and Team Solo only.
Args: fmt: "json" for the raw records, "csv" for one CSV block per table.
Your data is yours; this is the door out. It reads the local SQLite database and returns it inline — nothing is uploaded.
Inputs: fmt.
get_historyA
Return the full time-series history so Claude can surface trends.
Returns daily check-in snapshots, revenue by day, and calendar load over the requested window. Useful for building sparklines, comparing weeks, or asking "am I actually getting better?"
Free tier returns the last 7 days; Pro returns everything.
Inputs: days.
ingest_calendar_eventA
Ingest a calendar event. kind ∈ {sales, deep_work, admin, personal}.
As with revenue: manual entry is free, connector-sourced ingest is Pro.
Inputs: kind (required), end_iso (required), start_iso (required), title, +2 more.
ingest_revenue_eventA
Record a revenue event. Amounts in the user's home currency.
Typing one in by hand (source="manual") is free. Automated ingest from an upstream connector — Stripe, QuickBooks, a wearable bridge — is a Pro feature.
Inputs: amount (required), meta, iso_ts, source.
license_statusA
Show the active Tempo plan and what it unlocks.
Call this whenever the user asks what they're paying for, why a tool refused, or how much history they can see.
Inputs: no named parameters.
log_checkinA
Log a morning (or moment-in-time) founder check-in.
Args: mood: 1..10, subjective mood energy: 1..10, subjective energy stress: 1..10, subjective stress sleep_hours: hours of sleep last night (can be fractional) note: optional free text
Inputs: mood (required), energy (required), stress (required), sleep_hours (required), +1 more.
recommend_actionsB
Return prescriptive business actions based on current Tempo state.
These are the decisions that separate Tempo from a wellness tracker: price raises, meeting declines, rest days, deep-work blocks.
Inputs: window_days.
tempo_scoreA
Compute the current Tempo score and supporting metrics.
Returns tempo (0..100), capacity, load, RpRH (revenue per recovery hour) and its 7-day delta, plus human-readable signals and alerts.
Free tier scores over the last 7 days; Pro over any window.
Inputs: window_days.
weekly_digestA
Generate a human-readable weekly digest Claude can read aloud or send.
Compares this week vs last week on every key metric: Tempo score, capacity, load, RpRH, meeting hours, deep work hours, revenue. Returns a narrative summary string plus structured deltas.
Week-over-week needs 14 days of history, so the comparison is Pro. Free tier gets this week's digest on its own.
Inputs: no named parameters.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 11 tools
Each tool has a clearly distinct purpose: scoring, check-in logging, revenue/calendar ingestion, recommendations, action application, history, digest, license management, and export. No two tools overlap in function, and descriptions make boundaries explicit.
Most tools follow a consistent verb_noun pattern (log_checkin, ingest_revenue_event, recommend_actions, apply_action), but a few like tempo_score, weekly_digest, and license_status are noun_phrases, slightly breaking the pattern. Still, all names are snake_case and readable.
With 11 tools, the count is well within the ideal 3-15 range and each tool earns its place in the workflow—from data ingestion to scoring, recommendations, and administration. No redundancy or bloat.
The tool set covers the full domain lifecycle: input (check-in, revenue, calendar), analysis (tempo_score, recommend_actions), output (get_history, weekly_digest), execution (apply_action), and administrative functions (license_status, activate_license, export_data). Missing update/delete operations are not critical for this append-heavy logging use case.
Maintenance
Related MCP Connectors
WHOOP recovery, strain, sleep and workouts in Claude via official WHOOP OAuth. Free, open source.
Apple Health training load, recovery, HRV and workout detail for Claude, ChatGPT and any MCP client.
N-of-1 health experiments in your AI assistant: Oura/RescueTime data, protocols, workouts.
Not another dashboard. A wealth analyst for every asset a bank can't sync, inside Claude.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceIntegrates WHOOP biometric data into Claude and other MCP-compatible applications, providing access to sleep analysis, recovery metrics, strain tracking, and biological age data through natural language queries.13MIT
- AlicenseAqualityDmaintenanceProvides Claude with real-time access to local health data including sleep, recovery, strain, and workouts from WHOOP and Apple Health, enabling informed context-aware interactions.61MIT
- AlicenseNot gradedqualityDmaintenanceConnects your Whoop health data to Claude, enabling natural language queries about recovery, sleep, strain, and workouts.MIT
- FlicenseNot gradedqualityAmaintenanceEnables read-only access to WHOOP recovery, sleep, HRV, strain and workout data through 16 tools, 4 resources and 5 prompt templates for Claude Desktop, Claude Code, Codex and GitHub Copilot. Lets users ask natural-language questions about their trends, personal baselines, period comparisons and sleep debt without exporting data themselves.14 npm-