Document Intelligence that runs on your own infrastructure

We design interactions to mirror conversational habits. Synapse supports natural language queries and we offer quick-access integrations that surface answers inside tools your teams already use, reducing the need to navigate a new interface.

Synapse answers from the documents you connect rather than from general training data. Every response is returned with a citation to the source passage it came from, so a reviewer can check any claim against the original document instead of taking it on trust. Where the connected documents do not contain an answer, Synapse reports that rather than composing one.

That depends on the deployment you choose, and it is your choice rather than ours. On-premise, both the document corpus and the search index stay inside your own boundary — on servers you own, or in your own AWS, Azure, or GCP tenant, where the subscription, the storage, and the region remain yours. Nothing has to cross the network edge for the platform to work. On SaaS, the platform runs in certified data centres we operate, with tenant isolation enforced at both the application and the storage layer, and SaaS Enterprise adds a dedicated isolated tenant with a contracted SLA.

On an on-premise deployment, model access is yours to configure. You can point Synapse at a model running locally inside your own network, or at a provider you already hold an agreement with, under your own terms and your own key. We do not require you to route documents through a model we control, which is what makes a no-egress deployment possible at all.

Yes. On your own infrastructure with a local model, Synapse needs no route to the outside world to do its job — and neither the licence nor the update path requires one either. The platform does not phone home to stay licensed, and updates are applied on your own schedule rather than fetched from us. A machine that never sees the internet is a supported deployment, not a workaround.

Client-encrypted data is protected through a key management service, and every key access is recorded — who requested it and why the key was released — so key use is auditable rather than assumed. Documents you designate as critical are encrypted on the client before they are transmitted at all, and no key for those is stored on our side. We cannot read them, and we cannot be compelled to.

Access is role-based and applies to the document itself rather than to the folder it happens to sit in. You control who can view, update, approve, and publish each class of record, and the same model covers branch and head-office separation as well as time-boxed access for an external auditor. Every view, change, and approval is written to an access log, so the audit trail is a byproduct of normal use rather than something assembled afterwards.

You keep it, and you can get it out. On-premise the question largely answers itself — the documents and the index already sit on storage you own, inside your own boundary, so ending a licence does not move them anywhere. In either deployment, records and their metadata export in non-proprietary formats, JSON and CSV, so there is no vendor container to unpick and nothing that only Synapse can read.

Not yet. We are working towards SOC 2 certification and we do not claim it as though it were complete — we are happy to share where the programme currently stands. Separately, we follow GDPR practice across data subject rights, retention limits, records of processing, and breach handling, and that applies to every deployment mode rather than only the hosted one.

Our solutions team handles deployment end-to-end, from initial configuration to final handover. A typical deployment completes within days, and we provide ongoing support so your team is never left to manage infrastructure alone.

Yes. Synapse is built to ingest from fragmented sources - shared drives, email attachments, scanned PDFs, legacy DMS exports, and physical archives via bulk digitization. The consolidation step normalizes everything into a single searchable layer.

Yes. Updated documents are ingested on your defined cadence. Synapse maintains version lineage so that teams always reference the current approved revision, with full audit trails showing who changed what and when.

Yes. Documents can be routed through preparation, review, correction, approval, publishing, and archival stages based on your organization's roles and rules.

Yes. Different workflows can be created for policies, circulars, contracts, HR files, loan documents, procurement files, compliance records, and branch-level documents.

The document can be revised and resubmitted while preserving comments, previous versions, reviewer notes, and the full audit trail.

Yes. Teams can track pending approvals, delayed tasks, assigned owners, document status, and upcoming renewal or expiry dates.

We offer a structured pilot program that lets you validate Synapse against your actual document corpus and operating context before a full commitment, so you can build the internal case with real results.

It depends on how you deploy. On-premise is a flat annual licence with unlimited seats, running on your own infrastructure and your own model access, so nothing is metered at all. SaaS and SaaS Enterprise are priced per seat on graduated bands, and do meter three things: AI credits, OCR pages, and storage. Every one of those rates is published on our pricing page rather than quoted case by case, so the variable component is something you can model before you commit, not a surprise on the invoice.