Privacy

What Sploink is

Sploink is a behavioral telemetry layer for AI agent sessions. It captures what your agent does at the execution level — shell commands, file edits, git state, test results — so you can see whether it is converging on a solution or stuck in a loop.

What we collect

When you run a session, Sploink captures the following behavioral signals:

• Shell commands and their exit codes
• File edit events: path, lines changed, diff size in bytes
• Git state: diff size, commit events
• Test results: pass count, fail count, total count
• Build outputs: success or failure, duration
• Process spawns: command name, duration

What we never collect

We do not collect the contents of your files, environment variables, credentials, API keys, passwords, or any data that could identify what you are building. We capture behavioral shape, not substance.

• File contents
• Environment variables or secrets
• API keys, tokens, or credentials
• Command output beyond exit code
• Clipboard contents
• Network traffic

Where your data goes

Session data is written locally first, to ~/.sploink/sessions/. Every session is stored there regardless of your sync settings.

By default, session telemetry is also sent to Sploink servers. This is how the product improves over time: more sessions produce better convergence models, better intervention recommendations, and better agent rankings. Your sessions contribute to a shared model that makes every future session smarter — including yours.

Opting out

You can disable cloud sync at any time. Sessions will still be captured and stored locally — you just won't contribute to or benefit from the shared model.

Data retention

Local session data persists until you delete it. Cloud session data is retained for 90 days by default. To delete your cloud data, contact privacy@sploink.ai.

Contact

Questions or concerns? Email privacy@sploink.ai. We will respond within 48 hours.