Open-source AI for small landlords
Tickler reads your lease, remembers the important dates hidden inside it, and shows you the exact sentence behind every answer. It uses open-weight models, its reminders survive restarts, and your documents stay on your machine.
How it works
Drop in a PDF. Tickler reads it and shows each important date with the sentence it came from, the page number, and how sure it is.
See something wrong? Fix it in one click. Your corrections are saved, so the answers get better over time.
Reminders arrive 90, 30 and 7 days before each date. They keep working even after a restart. Letters come to you already written, and you decide whether to send them.
Why Tickler
Every answer shows the exact sentence and page from your lease. Tickler checks each quote against the document automatically. If something is not written in the lease, it says so.
“shall commence on May 1, 2027 and shall continue through October 31, 2027”
Reminders are built on Temporal workflows. In our demo we stop the worker program, and every reminder still arrives on time.
Tickler uses open-weight models, so your lease never travels to a closed service.
no closed servicesLetters are written using the sentences from your lease, and they always wait for your approval.
Your fixes and preferences are saved in the database. The more you use it, the better it knows your properties.
Six tools are available through the Model Context Protocol, so you can connect Tickler to the AI tools you already use.
npx tickler-mcp # six Cypher-backed tools, stdioThe principle
The software must never forget a date. The human must make the decisions.Tickler keeps both of these promises.
Free, open source, and your data stays on your machine.