Grounded in the right documents

Know the answer.
Know the source.

Meet RightSeat, an AI aviation knowledge companion built to help pilots prepare with aircraft-specific answers they can check against the actual manuals.

Watch the demonstration Explore RightSeat

Aircraft-specific Page citations Evidence first

Your exact aircraftDocuments that apply to you

Answers with sourcesCheck the original page

Actual app screenshot

Built around your aircraftNot a generic answer about every plane

Source pages, not just claimsSee where the answer came from

Designed to know its limitsInsufficient evidence should mean no guess

Why RightSeat exists

Because sounding right isn't the same as being right.

Aviation procedures depend on the exact airplane, its generation, and its approved documents. An AI answer can sound convincing even when it applies to a different aircraft.

RightSeat starts with a simpler idea: before trusting an answer, a pilot should be able to see the source.

How it works

A question is easy.
Finding the right evidence is harder.

RightSeat is designed to make trusted aviation references easier to use without losing the connection to the original documents.

Set up your aircraft

Choose your flight school or register an aircraft directly. RightSeat connects your questions to the reference documents relevant to that setup.

Ask naturally

Type or speak an aviation question, from an aircraft procedure to an FAA rule or a topic you're reviewing for a checkride.

Check the source

Follow page-level citations back to the original document. RightSeat is designed to flag answers it cannot support instead of guessing.

RightSeat is currently configured for multiple generations of Cirrus SR20 and SR22 aircraft. The configuration tested so far is the SR20 G7+.

Inside the app

A simpler way to work through difficult aviation questions.

Flying involves a lot of information: FAA handbooks, regulations, aircraft operating manuals, and flight-school procedures. RightSeat brings those references into one place to help you prepare.

Ask about your airplane

Find aircraft-specific information without searching through unrelated models or generations.

Go straight to the original reference

Review the cited page to understand the source, rather than relying on an unsupported summary.

Prepare for the oral and checkride

When you get stuck on a mock checkride question, use RightSeat to look up the topic and study the relevant source document.

Screenshots

Take a look inside RightSeat.

From choosing a question to speaking with RightSeat, practicing airport scenarios, and reviewing your document library.

RightSeat home screen with aircraft N546MY, question input and talk button — 01 / 05Ask RightSeat

RightSeat sign-in screen with email and password fields — 02 / 05Your account

RightSeat voice assistant listening screen with aviation question suggestions — 03 / 05Talk it through

RightSeat What Would You Do practice question about taxiway runway markings — 04 / 05Practice scenarios

RightSeat Library screen displaying knowledge-health summary and aviation source documents — 05 / 05Document library

Swipe sideways to explore all five screens.

See it in action

Watch the RightSeat demonstration.

See how the app is designed to connect aviation questions to aircraft-specific documents and show the evidence behind the answer.

Official RightSeat video

Having trouble viewing the video? Watch the RightSeat demonstration on YouTube.

Under the hood

Not just an AI chatbot.
An entire system built around evidence.

RightSeat runs its own AI backend on a Mac Studio at home. Rather than sending every question to a paid frontier model, its documents, retrieval system, local language model, and verification pipeline work together on self-hosted infrastructure.

27 source documents, nearly 15,000 passages

Apple Vision OCR processes scanned material. PostgreSQL and pgvector help retrieve the most relevant passages for the selected aircraft and flight school.

A local 27-billion-parameter model

Qwen 3.8-27B, quantized to 8 bits, runs via LM Studio on a Mac Studio M5 Ultra with 96GB memory, behind a FortiGate 51G firewall.

Verification, not blind confidence

A separate verifier checks generated claims against cited passages. The system is designed to acknowledge insufficient evidence rather than invent an answer.

A custom-built mobile and backend stack

Python and FastAPI power the server, while TypeScript, React Native, and Expo power the mobile app. The local approach avoids per-question commercial AI inference charges.

How RightSeat works
From approved aviation documents to source-backed answers

Diagram of RightSeat's document ingestion, aircraft-filtered retrieval, secure local model, and answer verification pipeline

What's next

Preparing for takeoff.

RightSeat's initial testing focuses on the Cirrus SR20 G7+. The goal is to bring aircraft-specific, source-backed preparation to more Cirrus pilots and training centers, and eventually to a wider general aviation community.

Coming soon to theApp Store

Coming soon toGoogle Play