Air-Gapped or Private Cloud
Deploy on isolated on-premise hardware or within your own private cloud. No external APIs, no shared services — your data never leaves your infrastructure.
Limited Pilot Programme — Now Open
Your practice’s project knowledge, always at hand.
Context Grid lets architects and design teams search drawings, specifications, reports, schedules and project correspondence in plain English — without sending a single file outside your environment.
Built for architecture practices and design consultancies working with sensitive client and project data.
Hardware cost only during pilot
On-premise or private cloud
No third-party services. No data exposure.
The Problem
Design teams shouldn’t waste hours searching for the right document.
Your project knowledge already exists. The problem is finding it.
×
“Which drawing supersedes this one?”
Three emails and half a day later — still not certain which issued version is current.
×
Searching across multiple live projects simultaneously
Different folders, different naming conventions, different team members on each one.
×
New team members take weeks to learn a project
The real knowledge is spread across drawings, reports and email threads nobody can find.
×
Client-sensitive documents with no clear status
Superseded issued drawings mixed in with current ones. A liability risk every time.
×
Cloud tools mean client information leaves the building
That’s not acceptable for many clients and project types — and it shouldn’t have to be a compromise.
The Solution
Ask questions about any project. Get answers from your own files.
Context Grid connects to your existing project folders and file shares. No restructuring, no migration, no cloud upload. Your team searches across drawings, specifications, notes and correspondence the way they actually think — and gets answers with exact source references.
✓No folder digging.
✓No SharePoint migration.
✓No external upload.
✓Always shows the current revision.
Context Grid
Is there a planning condition restricting working hours on this project?
Yes — Condition 9
restricts
construction works to 08:00–18:00 Monday to Friday and
08:00–13:00
on Saturdays. No works permitted on Sundays or bank holidays.
[1]
23-04521-FUL-Decision.pdf · p.3
[2]
Planning-Conditions-Schedule.docx
What brick was specified for the north elevation?
Ibstock Ravenhead Antique
Buff,
laid in Flemish bond with a 10mm natural sand mortar joint. See the External
Materials Schedule and elevation drawing GA-201 Rev C.
[1]
External-Materials-Schedule-Rev2.pdf
[2] GA-201-RevC.pdf
Ask anything about your project
files…
How It Works
Built for real design environments.
No migration project. No change to your folder structure. Context Grid uses a local AI model to index and search your documents — running entirely within your own environment — on-premise or private cloud.
1
Point at your file share
Connect to your existing network share or local data folder. No restructuring required. Context Grid works with the folder structure you already have.
2
Indexing begins automatically
Context Grid works through your project archive, extracting document content, metadata and revision status. File changes are detected and re-indexed automatically.
3
Start querying in natural language
Open the browser interface from inside your network. Ask questions in plain English. Receive answers backed by your actual documents, with precise source citations and revision status.
Superseded revisions are detected and excluded from normal search results by default — so your team always gets the current picture, with the ability to look back when needed.
Supported Formats
What Context Grid reads.
Works with the file types already on your server. No conversion, no pre-processing, no manual tagging.
Full text, page-by-page with precise citation
DOCX
Specs, schedules, reports, correspondence
XLSX
Room data sheets, finishes schedules, door and window schedules
DWG
Text layers, title block data, drawing number & revision
TXT / MD
Correspondence, meeting notes, design intent documents
Archives
Legacy project archives and deep folder structures
Automatic metadata extraction
Drawing numbers, revision levels, project codes and document status are extracted automatically from filenames and folder structures. No manual tagging required.
Revision awareness built in
Context Grid detects revision patterns in filenames and document metadata. Superseded revisions are flagged and excluded from default results — your team always gets the current document.
Continuous file watching
When a file changes, the index updates automatically. No scheduled jobs, no manual re-indexing. The knowledge base stays current without any intervention.
Why Context Grid
Why architecture practices choose Context Grid.
Understands Revision-Controlled Documents
Recognises revisions, superseded files and structured project metadata. Always the current picture, full history available when required.
Works With Existing Folder Structures
No migration project required. Keep using your current shared drives exactly as they are. Context Grid adapts to your structure.
Natural Language Search
Ask questions the way your team actually thinks. No Boolean operators, no exact filename recall. The system finds the document.
Designed for Architecture and Design Environments
Built for practices and consultancies where client confidentiality and project sensitivity are non-negotiable. Ideal for:
Architecture
Practices
Multidisciplinary
Consultancies
Interior
Design
Urban
Design & Planning
Heritage
& Conservation
Public
Sector Projects
Design
Management
Data Sovereignty
Your data stays yours.
Context Grid was designed for firms where client confidentiality is non-negotiable. Everything runs locally. No internet connection is required after setup. You remain in full control of your project data at all times.
Zero network egress
Context Grid operates with no outbound connections. Your documents are not indexed by any third party — not during setup, not during operation, not ever.
No third-party model exposure
The retrieval model runs locally. Documents are never sent to OpenAI, Google, Anthropic, or any other external service.
Air-gap compatible
Fully operational in isolated network environments. Appropriate for sensitive public sector, government and heritage projects where outbound connectivity is not permitted.
No licensing callbacks or telemetry
No outbound connections for licensing verification, usage reporting, or analytics. The system operates entirely within your network boundary.
Local language model
Local vector database
Local document storage
Private deployment only
Pilot Programme
Become a Pioneer Customer.
We’re currently working with a small number of architecture practices and design consultancies to shape the next generation of Context Grid.
Pioneer customers receive
- +Full on-premise deployment
- +Direct support and configuration
- +Feature influence and feedback access
- +Hardware cost only during pilot phase
In return, we ask for
- —Honest operational feedback
- —Real-world usage insights
- —A willingness to help shape the platform
Request a Pioneer Deployment
Tell us about your practice and we’ll be in touch within 2 business days.
Request received.
We’ll be in touch within 2 business days.
Something went wrong. Please try again, or email us directly.
Stop searching
folders.
Start querying knowledge.
Context Grid runs on your own infrastructure — on-premise or private cloud — indexes your existing project documents, and gives your team answers in seconds, with no data ever reaching external services.
Request a Pioneer Deployment →Got a question first? Use the form above ↑