AI for FileMaker. Runs on your machine.

Single-file executables. No installation, no dependencies, no cloud accounts. Download, run, call from FileMaker via Insert from URL. That’s it.

Every tool runs on localhost. Your audio, your documents, your data — none of it leaves your network. No per-request billing. No API rate limits. No third-party terms of service to worry about.

Why Local Matters

Compliance without paperwork. No data processing agreements. No vendor security reviews. No explaining to your compliance team why client audio is being sent to a server in Virginia. The data never leaves the machine.

No recurring costs. No per-minute transcription fees. No per-page extraction charges. No surprise bills when usage spikes. Pay once, run forever.

No dependencies. No API keys to rotate. No vendor outages to wait out. No breaking changes when a provider updates their endpoint. Your tools work whether you have internet or not.

Fast. Local inference on Apple Silicon is fast. Transcription runs at multiple times real-time speed. Extraction completes in seconds per document. No network round-trip, no queue, no waiting.

Transcribr

Speech-to-text for FileMaker. Record a meeting, dictate a note, process an archive of recordings — get searchable transcripts attached to your records.

• Four modes: single file, batch processing, watch folder, real-time streaming
• Five models: choose your speed/accuracy trade-off per request
• Any audio format: MP3, WAV, M4A, OGG, FLAC and more — converted automatically
• Core ML accelerated on Apple Silicon
• Clean JSON response — parse with native FileMaker functions, no plugins

Transcribr Free is available now for Mac.

Download Transcribr Free   See full details →

Extractr

Structured data from any document. Feed it a PDF, a scan, or an image — tell it what fields you need — get clean JSON back into FileMaker.

• Schema-driven: you define the output structure, Extractr fills it in
• Any document type: PDFs, scanned documents, images, forms
• Batch processing: point at a folder, get structured data for every file
• Pre-built schemas for common document types (Companies House filings, invoices, medical reports)
• Same architecture: single file, localhost, Insert from URL, JSON response

Coming soon.

Coming Next

Semantic search and RAG for FileMaker. Ask questions in plain English across your entire record set. We’ve already built this for clients — 900,000 records searched in 2 seconds. The productised version is in development.

How It Works

Every product follows the same pattern:

1. Download a single executable for your platform
2. Run it — a local server starts on localhost
3. Call it from FileMaker with Insert from URL
4. Parse the JSON response with native FileMaker functions

No plugins. No external software. No configuration files. If you can write a FileMaker script, you can use these tools.