Bimiq Logo
Bimiq
Pricing
ContactLog in

Platform

Ask the project. It has already read everything.

Bimiq's agent has read your whole project: the drawings, the specs, the submittals, the clashes, the RFIs and every decision the team logged. Ask it anything and it searches across all of them, shows you what it found and where, then drafts the answer. When the answer is an action, it takes it, with you confirming before anything is written.

It answers from your project, not from the internet.

Every answer is built from your project's own data, and the agent can reach all of it. Ask a question and it searches across whatever carries the answer.

Documents and specs

SpecificationsEvery indexed document

Drawings

The sheet registerTags, dimensions and notesWhere each one sits

Findings

Drawing QASubmittal verdictsClashesModel vs specDocument conflicts

Record

RFIsThe decision log

It searches, reads, and searches again, until it can answer.

The agent writes its own queries, reads what comes back, and keeps going, up to fourteen steps in a single question, across everything it can reach. Then it answers with the sources attached. It doesn't answer from memory, and when it couldn't find something, it says so rather than guessing.

“What's still outstanding on the Level 5 party walls?”

search_documents("party wall acoustic separation, Level 5")

09 21 16 (2.4) · 54 dB DnT,w minimum between dwellings0.90

get_decisions(tags: ["Level 5", "Acoustic"])

12 Feb · party walls upgraded to twin-stud SW-02 after acoustic review

get_drawing_issues(axis: "cross-sheet")

A-201 still shows single-stud SW-01, not SW-02

get_clash_issues(level: "Level 5")

2 critical · MECH ↔ STR, +16m Bay E4

A-201 still shows SW-01, but the team moved to SW-02 in February, and two clashes remain on the level. Every line above is a source you can open.

When the answer is an action, it takes it.

The same agent that reads your project can write to it, so an answer doesn't become a to-do list you work through by hand.

Draft an RFI

Ask it to raise an RFI and it writes one, precise and referenced, for you to review before it becomes real.

Raise an RFI from a finding

Straight from a drawing finding or a submittal comment, linked back so the finding shows the question was asked.

Log a decision

Tell it what the team settled and it records the decision, embedded so the next question can find it.

Nothing is written without you confirming, and nothing is sent.

With the Bridge open, it reaches the live model too.

Query and edit Revit

Ask what's on Level 3, or change a parameter across the doors that fail a rule. The agent runs the operation on the open model and shows you the result.

Read the Navisworks federation

Reach the federated model and its clash tests, so a question about coordination is answered from what is actually in the combined model.

The model never leaves your machine, and the agent confirms before it writes anything back. Without the Bridge, it works from the documents and tells you the model wasn't consulted, rather than answering with two-thirds of the picture.

See what it finds in your first set.

Upload a drawing set, specification or submittal and let Bimiq review it. Every finding links back to the exact sheet, clause or model element so you can verify it yourself.

14-day free trial. No card required.