Every team running automations eventually hits the same wall: someone asks "so how much is all this actually saving us?" and the honest answer is a shrug and a spreadsheet. Time Hive is our answer to that question, built so the number you give back is one you can defend.
Time Hive is a multi-tenant SaaS platform that records automation activity as an append-only ledger and derives reporting metrics — chiefly time saved and tasks completed — from those logged facts. It started as an internal tool for measuring the hours our own automations claw back, and it's now a product anyone can plug their workflows into. The principle that runs through the whole thing is short enough to put on a sticker: facts are logged, metrics are derived.
The problem: efficiency you can't prove
Automation ROI is usually self-reported, and self-reported numbers are soft. If a dashboard lets you type "we saved 200 hours" into a box, that figure is a guess wearing a suit. The moment a client, a CFO or a board member leans on it, it folds.
Time Hive's design goal is the opposite of soft. There is no field anywhere in the product for entering "minutes saved" by hand. Every headline figure is computed from two things the platform can actually point to: the events your automations emit, and the per-task assumptions you declared up front. That makes the totals audit-friendly — the numbers you report are backed by data that can't be quietly fudged after the fact.
Who it's for
Time Hive is built for the people who have to show their working: automation engineers, agencies and consultants proving the value of what they shipped, and ops teams justifying the tools they run. If you've ever had to make a renewal case for an automation and reached for a spreadsheet of estimates, this is the tool you wanted.
Feature 1 — An append-only ledger
Everything starts with the ledger. Each automation run and each output is written once, as an immutable record, and never edited or deleted. Corrections happen by exclusion or recalculation, not by reaching back and rewriting history. That's what makes a reported number defensible: the audit trail is the data, and the data only ever grows.
Feature 2 — Platform-agnostic ingestion
Time Hive doesn't care how your automation is built. Anything that can send a webhook can feed it: n8n, Make, Zapier, a Python script, a cron job, your own code. Each automation gets its own integration — a live ingestion endpoint with a per-integration ingest token and webhook URL. You add one HTTP call to a workflow you already have, and the events start flowing.
Feature 3 — Facts logged, metrics derived
This is the heart of it. Time Hive never stores a metric — it stores events, then derives the metrics on the way out through SQL views (time_saved_rollup, outputs_rollup and friends). Raw run and output events go in at the bottom; rolled-up totals come out the top. Because the totals are recomputed from the underlying facts, they're always consistent with the ledger and impossible to inflate by hand.
The events themselves are cheap and robust to ingest. A single call can batch many executions with a count field, and an idempotency_key means a webhook that fires twice is silently de-duplicated rather than double-counted — a feature, not a bug, when networks misbehave.
Feature 4 — Time saved, computed not claimed
So where does "hours saved" actually come from? Not a text box. Each task carries a declared assumption: how long it takes a human (manual_time_min) versus the automation (automated_time_min). Time Hive multiplies the difference by the number of confirmed run events. The assumption is visible, the run count is logged, and the product of the two is your defensible total.
Run that across a whole account and you get the dashboard Time Hive is built around: time saved by tag, run volume, error rates and rolling 7/30/90-day totals — each one shareable with a stakeholder via a read-only link. There's even an optional public ledger that tallies minutes saved across every project, for teams who want to show the number off rather than hide it.
How it works, end to end
The whole loop is three steps:
- Create a project and define a task with its manual-versus-automated time assumption.
- Send events by adding one webhook call to an automation you already run —
event_type, anidempotency_key, and thetask_idfor runs. - Track ROI as the dashboard recomputes time saved and tasks completed from the facts, in real time.
Under the hood it's a small, modern stack: a Next.js web app, a Hono API on Cloudflare Workers handling ingestion, Supabase (Postgres) with row-level security keeping every tenant's data isolated, and Stripe for billing. The free plan runs a single project with no credit card — enough to wire up your first automation and watch the counter move.
Why we built it
Time Hive is the clearest expression of the thing Swarm Labs cares about: we don't ship features, we hand back hours — and then we count them honestly. It's one of several internal tools we've turned into products, alongside SwarmGen, Apiary and Scout. If you'd like to read more about how we measure recovered time, the rest of the hive's notes are a good place to start, or see Time Hive on our applications page.