Your AI guesses.
InputLayer reasons.
Store facts. Write rules. InputLayer draws conclusions, keeps them up to date as things change, and can show exactly how it got every conclusion.
How it works
Give your AI conclusions it can trust
InputLayer figures out what's true from the facts and rules you give it. Your AI gets reliable, up-to-date conclusions instead of guessing, and every conclusion can be traced back to the facts behind it.
Thinks, not just stores
Your AI stores facts. InputLayer connects the dots and draws conclusions automatically.
Search + logic together
Find things by similarity, then filter by what's actually true. One query, both at once.
Shows its work
Every conclusion traces back to the facts and rules that produced it. Ask "why?" on any result.
Always up to date
When a fact changes, every affected conclusion updates in milliseconds. No waiting, no rebuilding.
Comparison
What you get that other tools don't do
Databases store data. InputLayer thinks about it, combining search, logic, and live updates in one place.
| Capability | Vector DBs | Graph DBs | SQL | InputLayer |
|---|---|---|---|---|
| Find by similarity | plugin | plugin | ||
| Follow connections | partial | |||
| Apply rules | partial | partial | ||
| Chain logic together | ||||
| Live updates | partial | |||
| Undo when facts change | recompute | recompute | ||
| Explain every conclusion | partial | partial |
Search + logic
Where can I fly that feels like a beach vacation?
You're in New York and want beach destinations, but only ones you can actually get to. First, InputLayer follows the flight network to figure out which cities are reachable (even with connecting flights). Then it ranks those cities by how well they match "beach vacation."
One query does both
Rules figure out where you can go. Search ranks by what you want. InputLayer does both in a single query. No stitching things together, no extra steps.
Open in StudioSmart undo
Cancel a route. Does the destination stay reachable?
Sydney is reachable from New York two ways, through Tokyo and through Dubai. Cancel the Dubai route, and most systems would just wipe the conclusion. But the Tokyo route still works.
InputLayer tracks both paths. It only removes a conclusion when every way to reach it is gone. Sydney stays reachable until both routes are cancelled. No wrong removals, no stale conclusions.
Open in StudioInstant updates
Add a route. Only the affected conclusions update.
Your flight network has 2,000 airports and hundreds of thousands of possible connections. Add one new route, say London to Dubai, and most systems recalculate everything from scratch. InputLayer only updates the conclusions that actually changed.
incremental update
full recompute
faster
Explainability
Ask "why?" and get the full reasoning chain
How does New York connect to Sydney? Ask .why and see every step: NY to London, London to Paris, Paris to Tokyo, Tokyo to Sydney. Or ask .why_notto see why São Paulo can't be reached. Every conclusion comes with the receipts.
of results fully explainable
Go deeper
The problems behind the features
Why Search Alone Isn't Enough for Your AI Agent
Finding similar items is useful, but it can't tell you which destinations are actually reachable. See how rules and search work together in one query.
Read the deep dive Extends: Smart undoWhy Delete Should Actually Delete
When a route gets cancelled, which destinations become unreachable? Getting the right conclusion is harder than it sounds, and getting it wrong can break everything downstream.
Read the deep diveOpen source. Run it yourself.
No account, no API key, no vendor lock-in. From first query to production in four steps.
docker run -p 8080:8080 ghcr.io/inputlayer/inputlayer