Use case · Journalists

Eyrie for journalists.

For investigative reporters, foreign correspondents, beat journalists, and freelancers. Interview transcripts that stay on your Mac. Source files that aren't in someone else's cloud. AI that helps you find the pattern across 50 interviews — and shows you which interview each fact came from.

The three things journalists need that other notes apps don't ship.

1. Source protection by default.

Notion stores your notes on AWS. Mem stores them on their servers. Bear stores them in a sandboxed database. None of these is what you want for source-protected reporting. Eyrie's notes are plain Markdown files on your Mac, end of story. No app server, no cloud sync (unless you opt in via iCloud / Dropbox). Subpoena resistance comes from physical location, not from a privacy-policy promise.

2. Verifiable AI for cross-interview pattern recognition.

You've done 40 interviews on a story. Three sources mentioned a specific name in passing. You don't remember which three. Eyrie's chat: "Find every interview where [name] is mentioned, even briefly." Sourced answer with clickable chips to each transcript. Verify by clicking. No invented attribution.

3. Plain text that outlives every app.

Twenty years from now, your reporting archive needs to be readable without booting up a specific app. Markdown in folders is the most durable digital format we have. Even if Eyrie disappears, every interview file you saved is still on disk, openable in any text editor.

The reporting workflow.

1. One file per source / interview. Name by date + interviewee surname. Frontmatter for: date, on/off-record, location, contact info. Tags for: #background, #named-source, #expert-source, #document.
2. Story files for each piece you're working. Drafts, outlines, follow-up questions. Smart Categories tag these as "project" so they live in the active sidebar.
3. "Threads" files for ongoing investigations. Persistent topics that span multiple stories.
4. AI as fact-check buddy. "Find every source I have on Claim X. List anyone who contradicts it." Eyrie searches your transcripts, returns sourced answers. Use as a starting point, never as the final word.
5. Encrypt the folder. macOS FileVault is on by default. For extra protection on specific projects, use an encrypted disk image as the vault folder.

A note on AI and source confidentiality.

This is the hard question. When you use any AI feature in any notes app, the relevant text is sent to the AI provider (Anthropic, OpenAI, Google). Their privacy policies vary, and their retention practices have changed over the years.

Two paths for sensitive reporting:

• Disable AI for that folder. Eyrie's AI features are opt-in per vault. For sensitive investigations, run Eyrie with AI off — you still get fast local search, backlinks, Smart Categories (which use local embeddings, no cloud).
• Local LLMs. For users who want AI without any cloud, point Eyrie at a locally-running model via Ollama. Slower than Claude, but truly air-gapped.

For non-sensitive reporting, the AI features pay off; for sources whose identity you'd protect under any circumstances, treat AI as you would a copyeditor — useful for some passes, off-limits for the raw material.

The other apps reporters use.

• Otter / Trint for transcription. Output as plain text into Eyrie folders.
• Zotero for PDF document storage and citations. Eyrie reads Zotero's note exports.
• Signal / SecureDrop for source comms. Eyrie is for the notes you take after, not the comms themselves.
• VeraCrypt / encrypted DMG for the vault folder on shared machines.