Automatic time tracker
Tockly vs Timely
Timely is the AI-timesheet product for agencies and consultancies: a private background tracker records app and window metadata, and its AI drafts billable time entries you approve, with a strict no-screenshots policy.
Tockly is work tracking that reads AI activity: every minute scored as hands-on, agent running, steering, or idle. Here is how the two compare, line by line.
At a glance
| Capability | Tockly | Timely |
|---|---|---|
| Classifies human vs AI-agent minutes | Every minute: hands-on, agent running, steering, or idle | AI files time to projects, not to human vs agent |
| Credits time supervising an AI agent | Reads agent activity directly and credits it as work | Sees a Claude window was open; cannot say who worked |
| Screenshots | Every 10 minutes, tied to the minute they capture | Never, by policy |
| Activity tracking | Input event counts only, never keys or content | App and window metadata, private to the user |
| App and site usage | Active app and window titles | Private timeline; managers see approved entries only |
| Invoicing from tracked time | Free for up to 3 clients, unlimited on Pro | Native invoices from tracked time, plus QuickBooks sync |
| Free tier | Full capture, unlimited activity history | None, 14-day trial |
| Entry price (annual) | Free. Pro $12, Team $16 per seat/mo | $9/seat/mo |
Timely details verified against its public pricing and documentation as of July 2026. Products change; check Timely's site for current numbers.
The minute Timely can't see
Timely's AI answers one question: which project should this block of time bill to. The nature of the time never enters the record. Since May 2026 its tracker recognizes Claude and Codex windows by title, but that is presence, not classification: it knows a Claude window was open, not whether the agent was working while you watched or you were steering it. And because the raw timeline is private and screenshot-free, the timesheet a client sees is still self-reported at the moment of approval.
Pricing
Tockly
- Personal: free. Full capture, invoicing for 3 clients, screenshots from this month and last.
- Pro: $12/seat/mo annual ($15 monthly). Unlimited clients, 12 months of screenshot receipts.
- Team: $16/seat/mo annual ($20 monthly). Org rollups, roles, per-person AI adoption.
Timely
- No free tier; 14-day trial.
- Starter: $9/seat/mo annual ($11 monthly), up to 5 users and 20 projects.
- Premium $16 annual ($20 monthly), Unlimited $22 ($28).
Which should you pick?
Pick Timely if the pain you are solving is timesheet drudgery on a trust-based team: its AI drafting from automatic capture is the most mature in the category, and its privacy stance keeps rollouts friction-free.
Pick Tockly if you need to know or prove how the hours happened. Timely's AI files time to the right project; it cannot say whether a human or an agent worked it, and its private-timeline design means nothing is verifiable. Tockly classifies every minute and keeps the receipts.
Frequently asked
Timely tracks Claude windows now. Does it measure AI work?
It recognizes Claude and Codex window titles so they categorize cleanly, the same way it recognizes any app. That is presence detection. It has no agent event feed, no token data, and no concept of agent running versus steering, so an afternoon of supervised agent work is just window-title metadata filed to a project. Tockly's classification comes from the agent's own activity, minute by minute.
What is the difference between Timely's AI and Tockly's AI awareness?
Timely's AI is a filing assistant: it learns your logging habits and drafts timesheet entries you approve, which genuinely kills timesheet drudgery. Tockly's AI awareness is the measurement itself: it changes what the record can say, adding who was driving each minute. One automates the paperwork about your time; the other upgrades what your time record contains.
Can Timely prove billed hours to a client?
Only as far as trust extends. Timely's raw timeline is private to each user and it takes no screenshots, so what reaches the client is whatever entries were approved. Tockly's record is built for scrutiny: per-minute classification with periodic screenshots, so a disputed line item resolves with evidence, including how much of the hour was agent work.
More comparisons
See your own human/agent split today.
Install the app, work a normal day, and the record builds itself, minute by minute.