Product
The pipeline, end to end.
Six visible stages turn a written brief into a checked, shareable decision. What you see is what runs.
Deliver — only now does an answer leave the stack.
Six stages
What runs, stage by stage — including what stops it.
Every stage names what it took in, what it produced, and where each value came from. It also names how it can fail, because a pipeline that can only succeed is not a pipeline you can trust.
Stage: Describe
Describe
A written brief in business language — no form, no modelling syntax.
- In
- What you type
- Out
- The brief, and the kind of decision it is
What this stage holds
- Brief
- One van, five drop-offs, back to the depot. Which order is shortest?you wrote this
- Detected
- Routingdetected
- Confidence in that reading
- shown, not hiddendetected
How it can fail
Failure mode: If the brief does not describe a decision, Mongeflow says so rather than guessing at one.
STEP 01 • DESCRIBE
A brief, not a form
The intake is a written brief in business language. It is checked before anything runs, and the type of decision is detected with a reported confidence.
STEPS 02–03 • UNDERSTAND · CONFIRM
Understood, then confirmed
What it understood
The goal, the data it will use, and your business rules are shown back as an editable brief. Every value carries its source — what you see is exactly what gets used.
Missing details, named
Before anything is recommended, the brief is checked for completeness. Anything still missing is listed plainly — no silent assumptions.
STEP 04 • RECOMMEND
A recommendation with its impact
You get a clear recommendation with its expected impact, the confidence behind it, and the alternatives — a fast first answer, then a thorough checked result.
STEPS 05–06 • VERIFY · SHARE
Checked, then shared
Checked before shown
Every recommendation passes a full set of checks before you see it. It is marked verified only when all of them pass — and the trail of how it got there travels with the answer.
From decision to report
Recommendation cards, scenario comparisons, report and reproducibility formats, API reads, and signed webhooks carry the decision to wherever the work happens.
THE METHOD
Proven rigor, in plain language.
The decision methods specialist teams rely on, reachable in your own words — with the working shown.
THE API
An API for what you decided
Keys are mf_live_xxxxxxxxxxxxxxxxx and mf_test_xxxxxxxxxxxxxxxxx bearers. Rate limits scale by plan — 30, 120, or 600 requests per minute, with X-RateLimit headers on every response. The full surface is documented in OpenAPI 3.1 at /api/v1/openapi.json.
# Every webhook is signed
X-Mongeflow-Signature: sha256=…
# HMAC-SHA256 over the payload — verify before trusting
# Signed event types include:
analysis.solved · the decision is ready
export.ready · the report is ready
Questions
What teams ask first
Can I see what it understood?
Yes. The decision brief shows the goal, the data, and your business rules, with a source on every value, and it is editable before anything is recommended.
How does Mongeflow reach an answer?
It uses proven, rigorous decision methods — a fast first pass for an early answer, then a thorough checked run. The first pass is labelled as a draft.
What does “verified” mean?
A recommendation is marked verified only when it passes every check after it is prepared. The full set of checks travels with the decision.
Can I edit what it understood?
Yes — the decision brief is editable, and the checks re-run after every change.
How is my data handled?
Organizations are isolated at the database layer with row-level security. Our infrastructure providers maintain SOC 2 Type II certification.
What are the plan limits?
Decisions per month, seats, report formats, and API rate limits scale by plan — see pricing for the exact numbers.