Integration Engineers
The interface that passes on 14 test patients
will fail on 14,000 real ones.
You are the person who keeps clinical data flowing. ADTs from registration. ORUs from the lab. RDEs to the pharmacy. When an interface fails, you are the one who gets the call — and you are the one who fixes it at 2am so a patient can get their care.
50 ADTs: 4 hours by hand. 30 seconds with Post.
Clinically coherent demographics, vendor-specific field patterns, temporal coherence validated — in a single command.
You have done this long enough to know that hand-built test messages do not cover the edge cases that production throws at you. The 99-year-old patient. The hyphenated last name. The Z-segment your vendor added last quarter without telling anyone. You build 14 test ADTs and hope you covered enough. You deploy. You wait.
That waiting is the problem Pidgeon was built to solve.
What Go-Live Testing Looks Like Today
You open a text editor. You type MSH, pipe, caret, tilde, backslash, ampersand. You build PID segments by hand, checking field lengths against the spec. You create 10 or 20 test patients — all named Test Patient, all born in 1980, all with the same insurance. You send them through. Everything passes.
Then you go live.
And the first real patient has an apostrophe in their last name and your downstream system rejects the ADT. Your 14 test patients did not cover the edge case your vendor sends in production every day.
What Changes With Post
$ pidgeon generate ADT^A01 --count 50 --vendor epic
Generated 50 HL7 v2.3 messages → ./output/
✓ Clinically coherent demographics
✓ Epic-specific field patterns
✓ Temporal coherence validated
278 trigger events. 110 segment definitions. v2.3.1 through v2.8. Messages that match what your vendor actually sends — not what the spec says they should send.
Apostrophes, extended character sets, Z-segments, vendor-specific field population patterns. Every edge case you have ever debugged at 2am is in the generation engine.
$ pidgeon validate --file adt_batch.hl7 --mode compatibility
47/50 passed. 3 failures:
PID.5: Name exceeds 48-char vendor limit
(line 12, 34, 41)
Validate against real-world rules, not just the spec. Compatibility mode accepts the variance your vendor actually sends.
Everything the spec forgot to test
Post generates the messages that your text editor never would. The ones that cause Saturday morning phone calls.
Vendor-specific generation
Apply an --vendor epic, --vendor cerner, or --vendor meditech flag and get messages that match what those systems actually emit, not the idealized spec version.
Multi-version HL7 support
v2.3.1 through v2.8 on the same installation. Switch versions per message batch without reconfiguring anything. The right version for every interface.
Strict and compatibility modes
Strict mode enforces the full HL7 spec. Compatibility mode accepts the real-world variance your integration engine actually handles. Test both sides of the boundary.
Pre-Flight Check
50 clinical acceptance scenarios covering demographics, labs, pharmacy, radiology, and admissions. Run before every go-live. Know what will break before it reaches production.
Temporal coherence
Admission before discharge. Collection before result. Order before fill. All timestamps across related segments maintain logical ordering automatically.
Instant install
dotnet tool install pidgeon. No staging environment. No cloud dependency. No IT provisioning request. You are generating messages before your second coffee.
The engineer whose go-lives are clean.
Not because you are lucky. Not because you caught the error at the last minute. Because you tested with data that matches what your vendor actually sends, validated against real-world patterns, and found the failures on Thursday afternoon instead of Saturday morning.
Saturday morning, your phone does not ring.
You generate 500 vendor-specific scenarios in thirty seconds. Two fail on PID.5 field length. You fix the mapping before the go-live window opens. The issue never reaches production. The patient never waits.
$ dotnet tool install pidgeon
Tool 'pidgeon' installed successfully.
$ pidgeon generate ADT^A01 --count 10
✓ 10 messages written to ./output/
$ pidgeon validate --file ./output/ --mode compatibility
✓ 10/10 passed. 0 schema errors.
Generate your first message in 30 seconds.
One install command. No staging environment. No IT provisioning request. No cloud dependency. Start generating and validating clinical messages before your second coffee.
Built for how integration engineers actually work
Post is the CLI product designed for the engineer at the terminal. Everything else in the Pidgeon platform builds on top of it.
Post CLI
Open source. MPL 2.0. Every message type, every trigger event, every HL7 version. Generate and validate without any account, without any cloud, without any limits on core generation.
- pidgeon generate — all message types
- pidgeon validate — strict + compatibility modes
- pidgeon send — SFTP and local delivery
- pidgeon config analyze — vendor pattern detection
Post Pro
For engineers doing multi-environment go-lives. Workflow Wizard, directory diff, AI Triage, enhanced datasets, vendor profiles, and go-live checklists.
- Workflow Wizard — Admit → Lab → Discharge
- Directory Diff — field-level HTML reports
- AI Triage — BYOK "why did this fail?"
- Go-Live Checklists — DEV → UAT → PROD
Loft Observability
When the go-live is over and the interface is live, Loft watches everything flowing through in real time. Alerts on failures. Recommends fixes. Keeps your interfaces clean without you watching Mirth dashboards.
- Real-time validation on live traffic
- Email, Slack, Teams webhook alerts
- Per-channel throughput and error rates
- AI-guided remediation recommendations
Download the CLI. Generate your first message in 30 seconds.
$ dotnet tool install pidgeon
$ pidgeon generate ADT^A01 --count 10
Download the CLI. Generate Your First Message in 30 Seconds.
Post runs completely locally on your machine. No PHI leaves your network. No staging environment required. Enter your email to get the secure Desktop download and start generating vendor-specific test messages today.