Your agents can read your training history

Give an agent access to your training records and it can help you use them. Compare training blocks. Build a chart the app doesn't have. Find the sessions behind a trend. You can ask new questions of the work you've already logged.

What open and owned mean

Owned means you control a usable copy. You can keep it, back it up and choose which tools or people get access. Reading your exported records doesn't depend on keeping a tensr subscription.

Open means the format is documented. A tool can learn how the records are organized and what the fields mean. In tensr, your database is SQLite and routines use .workout.md. Thedata model androutine format give software a way to work with those files.

Open describes the format. Your personal records stay private unless you choose to share them. Publishing the instructions for reading a file doesn't publish its contents.

Bring your own agent

An agent with file access and a SQLite tool can inspect your export and query the records directly. A chat tool without database support can work from query results you provide. Reading the public docs teaches it the format; you still need to give it access to your data.

That means you can use an assistant you already work with. You don't have to wait for tensr to build a particular report or buy our planned coaching service to analyze your own export. Your chosen AI service may have its own costs and data-handling policies.

• “Compare my last two four-week blocks. Show sessions completed and recorded volume by movement.”
• “Chart peak force for this exercise over time. Separate different sensors and setups.”
• “Summarize my last month for my coach, with the sessions behind each observation.”
• “Turn this routine into a printable workout sheet.”

These questions can use the same files. A new tool can become useful to you without requiring you to start logging from scratch.

Try it with a copy

1. Export your database and keep the original unchanged.
2. Use the agent setup guide to introduce the formats.
3. Give your tool a copy of the export, read-only database access, or selected query results.
4. Ask one concrete question and have it show the records and calculations behind the answer.

For example:

Inspect the schema in this export. Count my sessions by week for the last eight weeks recorded. Show the date range, query and results. Flag missing dates rather than guessing. Leave the source file unchanged.

Access to records helps an agent check its answer; it doesn't make every conclusion correct. Comparisons still need consistent exercises, units and measurement conditions. A change in a chart alone doesn't explain its cause.

You choose where that analysis happens. A local tool can read a local copy; uploading to a hosted service shares the data with that service. Seedata ownership for tensr's free-feature policy and planned cloud privacy model.