Flag the suspect messages on the interfaces you already run.
Your dashboard shows the channel is up. Loft reads the messages themselves, and flags the ones that break the standard, like a malformed order, for your analyst to review. Message content never leaves your network.
- Install
- A desktop app for Windows x64 (signed) or an Apple Silicon Mac (not yet signed), or the CLI. It can also run as a background service.
- Where it runs
- Next to your interface engine. Message content never leaves your network.
- Connects to
- Watched folders, MLLP, TCP, HTTP, SFTP, SMB shares, database polling, and the Mirth API.
- Account
- A free Pidgeon account. Signing in needs internet.
- Free in the beta
- Monitoring is free during the beta.
Compared with what you use: Engine dashboards track connections and errors. Loft reads each message against the standard.
Not yet: Learning each sender's normal traffic and flagging drift is a paid feature, not part of the free beta. Loft flags messages for a person to review. It never changes a message on its own.
Messages · 24h
6,116
across 12 channels
Content findings
4
0.07% need review
Transport health
12 / 12
channels connected
Channel health
Last 24 hoursADT · Core registration
3,482 messages
ORU · Lab results
1,906 messages
RDE · Pharmacy orders
728 messages
Review queue
4 openOrder date is not a valid date
RDE · Pharmacy orders
Result value is not a valid number
ORU · Lab results
Know before anyone calls
Your engine can report a channel as running while a message on it breaks the standard.
What the engine dashboard shows
ORU_OUT · Lab results
Running
1,906 messages today · 0 errors
What Loft reads
Flagged for review
OBX-5 "9..8" is not a valid number
ORU^R01 · message 1,412 of 1,906
How Loft works
It reads your channels
Point Loft at the channels and folders you already have. It works with Mirth, another engine, several engines, or none.
It flags the odd ones
Loft checks each message against the standard. Anything suspect goes to your analyst's review queue.
It explains what it found
The built-in AI assistant says in plain English what looks wrong and what to check next. By default it runs on a local model.
Every catch leaves a record
Loft keeps a timestamped record of every finding: what it caught, where the message came from, and where it was going.
Try Loft on your own channels
Download the beta, or talk to us about setting it up on your channels together.
The other Pidgeon products
Post
Test what your vendor actually sends.
Visit Post →Flock
Realistic test patients, with no real patient data.
Visit Flock →Loft
You’re hereCatch bad messages on live channels.
Migrate
Move your data out, and count every record.
Visit Migrate →Conform
In PostTest FHIR endpoints against the guide you must meet.
Visit Conform →