plain text in, results out
Automations you can read.
Describe what you want done in plain English — Tockwork runs it on schedule with your own AI key. Every step readable, editable, and yours.
- 1.Search the web for news about [my industry] from the last 24 hours
- 2.Verify the top 3 stories from their sources
- 3.Write a short brief; skip anything already reported yesterday
Pricing shift: two of your competitors moved to usage-based plans this week — both confirmed on their sites.
Regulation: the EU published final AI guidance for agencies; nothing else new since yesterday's brief.
How it works
Write it
Numbered steps in plain English. That's the whole automation — no nodes, no flowcharts, no JSON.
- 1.Search the web for [topic] news
- 2.Verify the top 3 stories
- 3.Write a short brief
Tockwork runs it
On your schedule or webhook, with your own AI key. Every run is logged step by step.
- step 1 — 14 results
- step 2 — verified
- step 3 — brief written
You get the result
Delivered to your inbox, or pushed anywhere with a webhook.
Real automations, as written
These are the actual steps — what you write is what runs.
- 1.Open the pricing pages for [my competitors]
- 2.Compare against last week's saved prices
- 3.Email me only if something changed
- 1.When a new form lead arrives via webhook
- 2.Research the company, size, and funding
- 3.Score the lead and send it back to my CRM
- 1.Summarize this week's signups and churn
- 2.Format the summary as structured JSON
- 3.send_webhook to my n8n workflow
Your key. Your models. Your words.
Every run uses your own provider key — Anthropic, OpenAI, Google Gemini, OpenRouter, Mistral, or DeepSeek. You choose the model.
Keys are encrypted the moment you save them and never shown again — not in the app, not in logs, not to us.
Tockwork is a flat product. No token markups, no per-run charges, no metering anxiety.
Pricing
- Everything included — schedules, webhooks, tools, API
- Bring your own AI key, pay your provider directly
- Per-run cost shown after every run
Later: one flat price. Never a markup on your tokens.
Questions, answered
Do I need to know how to code?
No. Automations are written and edited as numbered plain-text steps, like a note to a capable assistant. If you can write a sentence, you can write an automation.
What does it cost to run?
Runs use your own AI provider key, so you pay your provider's token prices directly — a typical daily brief costs a few cents. Tockwork shows you the cost of every run.
Which AI models can I use?
Bring a key from Anthropic, OpenAI, Google Gemini, OpenRouter, Mistral, or DeepSeek — and through OpenRouter, effectively any model on the market.
Can it talk to my other tools?
Yes. Automations can be triggered by incoming webhooks, send outgoing webhooks (to n8n, Zapier, Slack, or anything with a URL), call custom HTTP tools you define, and everything is available over a REST API.
What happens when a run fails?
You get an alert with the failing step, every run is kept in history so you can see exactly what happened, and automations that keep failing are paused automatically so they don't burn tokens.