What Ostler knows about you

Creative Machines – the company that makes Ostler – never sees any of the data on this page. The Ostler app runs entirely on your own Mac. Nothing listed below is sent to us, to our servers, to any cloud, or to any third party. We have no access to your personal data, at any time, ever. There is no account to create, no login to our systems, no analytics pinging home. This page is simply a transparent list of what the software on your own machine imports, learns from, and deliberately never stores – so you can see exactly what is happening inside your own device.

What Ostler imports

Contact information Stored

Names, companies, job titles, LinkedIn URLs, email addresses, phone numbers. From LinkedIn, Facebook, Instagram, WhatsApp, iCloud.

Ostler does not import profile photos from any platform.

Connection dates Stored

When you connected on LinkedIn, friended on Facebook, followed on Instagram. These become timeline entries.

Calendar events Stored

Event titles, dates, locations, and attendee names from Google Calendar and iCloud Calendar.

Ostler does not store event descriptions or notes.

LinkedIn message metadata Stored

Who you messaged, when, how many messages, which direction. Per-conversation signal: reciprocity, frequency, date range.

Ostler does not store message content. This is a deliberate privacy choice.

Phone number cross-references Stored

Ostler matches phone numbers from WhatsApp and Twitter exports against your existing contacts. The match result (yes/no) is stored, not the full phone numbers.

Instagram close friends Stored

If someone is on your Instagram close friends list, they are tagged as a close connection in the graph.

What Ostler learns

Relationship warmth Computed

Based on conversation frequency, reciprocity, and recency. Ranges from "cold" to "close". Computed locally, never sent anywhere.

Cross-platform identity Computed

"John Smith" on LinkedIn and "Johnny S" on WhatsApp are matched as the same person using name similarity and identifier overlap.

Conversation facts Computed

If you record a conversation, the local AI extracts structured facts: who said what, decisions made, action items, topics discussed.

Coaching observations Computed

Over multiple conversations, the AI notices patterns in how you communicate: interruptions, talk-time balance, question-asking. Stored in a local SQLite database.

What Ostler never stores

Message content Never stored

The full text of your LinkedIn, WhatsApp, or Facebook messages is never imported or stored. Metadata only (who, when, how many).

Passwords or credentials Never stored

If a password is mentioned in a recorded conversation, it is redacted in the cleaned transcript. This is hardcoded and non-configurable.

Usage analytics Never collected

Ostler does not track what you search for, how often you use it, or which features you click. There is no analytics code in the application.

Anything in the cloud Never sent

No data is ever transmitted to any external server. The AI models run locally. Pull the ethernet cable and everything still works.

Full privacy policy