Design Partner Program
Modern Data Governance
Private AI

Would you tell your clients where their data actually goes?

Most AI tools send sensitive documents through a chain of vendors, APIs, subprocessors, and DPAs. Rendex shortens that chain by running core AI workflows inside a controlled private environment. Bringing the model to the data.

Boutique Investment Banks M&A Advisors Family Offices Private Equity Wealth Managers Hedge Funds Law Firms
Become a Design Partner View Trust Center
120B
Model
131K
Context Window
No
Model APIs
Governed
Data Path
Private
Deployment
Cited Answers
Audit Logs
Ready to deploy private AI on your trust boundary?

We're selecting design partners now. One practice group, real documents, deployment matched to your firm's requirements.

Become a Design Partner
Run AI inside a controlled data environment.
Rendex gives firms private AI workspaces where sensitive documents, records, permissions, and outputs stay inside governed data paths.
01

Set the trust boundary

Choose where Rendex runs: private hosted workspace, customer-controlled cloud, or confidential compute.

02

Govern the data paths

Connect documents, records, filings, contracts, notes, and internal knowledge through controlled ingestion and permission-aware retrieval.

03

Run private AI workflows

Ask questions, draft responses, compare files, extract facts, and generate cited outputs with permission checks and audit logs attached.

Private AI infrastructure for confidential work.
Rendex gives high-trust firms a private AI environment for sensitive documents, diligence, client work, investment research, internal knowledge, fund records, contracts, and business information. Instead of pushing confidential materials through a chain of external AI vendors and model APIs, Rendex brings the model to the data.
🔒

Deployment-matched data control

Document processing, retrieval, and inference run inside a private deployment model matched to your firm's requirements. Rendex avoids third-party model APIs and does not train on your data. Confidential Compute deployments can enforce stricter egress controls.

No vendor lock-in

Built on open-source models and standard hardware. No proprietary formats, no data hostage. Your documents remain yours export everything, anytime.

Built for confidential financial, legal, and business workflows

Workspace isolation, permission boundaries, page-level citations, audit trails, and guided workflows that match how high-trust firms actually work.

One interface. Ask for what you need.
Comparisons, clause tables, guided workflows, precedent search ask in plain English and get cited answers. No separate tools to learn.

Comparisons

“What changed between the March and June drafts of the MSA?” cited to both versions

Clause Tables

“Build a clause matrix across these 40 contracts (assignment, indemnity, termination)” outputs a table with every cell cited to document + page. Export as CSV (with citations).

Guided Workflows

Diligence checklists, deal chronologies, transaction memos, and precedent reviews structured outputs with citations, exportable as HTML

Precedent Search

“Do we have prior work product or transaction history relevant to this deal?” filtered by workspace

Bulk Ingest

Scale to 100K+ documents with batched processing, progress tracking, and pause/resume. Built for data rooms, diligence, and discovery.

Due Diligence

M&A and custom checklists. Scan workspace documents, verify items, export to CSV or DOCX for deal teams.

Compare & Memo

Side-by-side document comparison with significance filters. Draft memos with citations; export to DOCX.

Machine Vision OCR

Scanned PDFs are automatically detected and re-processed through an on-device vision-language model. Quality scored, no cloud OCR, no data leaves the enclave.

Core platform capabilities
Workspace summarization one-click, cached per workspace
Machine vision OCR scanned PDFs auto-detected and re-extracted by on-device vision-language model
Document comparison side-by-side, significance scoring
Draft memo citations, DOCX export, refine
DMS & data-room sync iManage, NetDocuments, Worldox, SharePoint, and local folder watch. Source-to-workspace mapping with scheduled sync (hourly to daily)
Full audit trail queries, logins, exports; CSV for admins
Folder watch auto-ingest drop files into a network folder, indexed automatically into the target workspace
3-tier document library Personal, Workspace, and Firm-level catalogs with promotion between tiers
Sandbox demo mode read-only evaluation environment for prospect demos, no setup required
Two-model strategy Fast model for everyday queries, Deep Analysis toggle for complex reasoning across confidential materials
Circuit breaker and concurrency controls automatic throttling protects GPU resources during peak usage
AI adoption has become a vendor-sprawl problem.
Every new AI tool can add another API, another model provider, another subprocessor list, another vendor DPA, and another security review. Rendex shortens that chain by running core AI workflows inside a controlled private environment with stronger data isolation, permissions, audit logging, and operational governance.

Hardware-attested privacy (Confidential Compute)

Confidential Compute deployments run H100-class private inference inside Azure confidential compute with hardware-backed attestation. Lower tiers use private hosted or customer-controlled cloud deployments matched to your trust boundary.

Restricted operator access

Confidential Compute deployments are designed for restricted operator access, customer-verifiable attestation, and customer-initiated support. Access controls and support procedures are defined during deployment.

Every Answer Cited

Click any citation to see the exact source text. Answers are citation-backed, and unsupported responses are flagged or refused.

Workspace Isolation & Permission Boundaries

Workspace-level permissions enforced at the query level. SSO via Microsoft Entra ID. Every access logged.

Private Cloud AI Hardware-attested privacy
Azure confidential compute.
Sealed by hardware. Verified by attestation.

For Confidential Compute deployments, Rendex can run H100-class private inference inside Azure confidential compute with hardware-backed attestation and stricter egress controls. Confidential Compute deployments are designed to keep sensitive workloads inside a hardware-isolated environment with customer-verifiable attestation. Same citations. Same workspace isolation. Same audit trail. Founded by Matthew Giordano for firms where data sovereignty isn’t optional, it’s the requirement.

30-Day Pilot accepting applications
Built for high-trust work. Auditable by anyone.
Every component is self-hosted, inspectable, and replaceable. The governance layer on top tuned to confidential financial, legal, and business workflows is what makes it Rendex.
Built for high-trust work
Domain-aware parsing
Section-boundary chunking, definitions detection, QC scan for missing standard clauses across contracts, deal documents, and filings.
RAG pipeline
Hybrid Qdrant + OpenSearch retrieval, citation grounding, optional cross-encoder rerank.
Diligence & review workflows
Draft generation + risk scan, tabular review, playbook runner, clause-level compare across deal, investment, and contract materials.
Governance
Per-document ACLs, query-time permission boundaries, append-only audit log, multi-tenant scoping, SSO via Entra ID.
Egress controls
Confidential Compute deployments can enforce stricter egress controls. Across deployments, traces and errors should avoid storing prompts and use hashed user identifiers where possible.
This is what you’re paying for.
Auditable stack
Models & serving
GPT-OSS 120B MoE on vLLM; BAAI/bge-base-en-v1.5 on HuggingFace TEI
Storage & search
Qdrant, OpenSearch, PostgreSQL, Valkey
Ingest
Apache Tika, Marker (PDF→markdown)
Observability
Langfuse (metadata-only LLM traces), GlitchTip (Sentry-compatible), Prometheus + Grafana
Runtime
Node/Express, Nginx, Docker Compose
Every component is open source, self-hosted, and independently inspectable.

The stack is inspectable. The way it’s put together for high-trust work is not.

Full deployment and security details in the Trust Center →
The largest firms build their own AI.
You shouldn't have to.
Large institutions hire 3–5 AI engineers, spend $800K–$1.5M per year, and take 6–12 months to deploy. Boutique and mid-market firms deserve the same capability without the same overhead.
What a large firm builds internally
3–5 engineers hired$800K–$1.5M/year in salary alone
6–12 months to deployA working system if the project doesn't stall
Ongoing maintenanceModel updates, security patching, infrastructure management
Built for one firmNo cross-firm improvements
Key-person riskIf the lead engineer leaves, the project stalls
What Rendex deploys for you
Pre-configured private AI workspaceDeployed in days, not months
Same hybrid search, citations, and workspace isolationEverything a large firm builds ready from day one
Quarterly updatesNew models, new features, security patches we handle it
Every improvement benefits every firmCross-platform improvements with every release
No AI team to recruit, manage, or retainOne private AI workspace. Deployment matched to your trust boundary. No AI team to recruit.
Pricing by trust boundary.
Rendex can start as a private hosted workspace and scale into customer-controlled cloud or confidential compute as the sensitivity of the work increases. The more sensitive the work, the more controlled the deployment.

Not every firm needs the same deployment model on day one. Rendex pricing is structured around how much control, isolation, governance, and compliance support the client requires.

Private Hosted
Customer-Controlled Cloud
Confidential Compute
Tier 1
Private Workspace
For firms that want private AI without third-party model APIs.
Pricing
Scoped per engagement
Best fit

Accounting firms, boutique advisors, smaller law firms, early design partners, and moderately sensitive workflows. Hosted in a dedicated GPU/cloud environment managed by Rendex.

Includes
  • Private model endpoint
  • Document ingestion
  • Vector search and RAG
  • Workflow packs
  • User permissions
  • Audit logs
  • No OpenAI or Anthropic API exposure
Enquire
Tier 3
Confidential Compute
For firms where sensitive data cannot leave a controlled trust boundary. Enterprise private AI.
Pricing
Scoped per engagement
Best fit

Family offices, law firms, investment banks, private equity firms, hedge funds, and regulated financial or legal workflows.

Includes
  • Everything in Customer-Controlled Cloud
  • Azure confidential compute target
  • H100-class private inference
  • Attestation story
  • Stricter egress controls
  • Audit-ready logging
  • Document-level permissions
  • Dedicated environment
  • Custom workflow packs
  • Compliance support
Enquire

Every engagement is scoped to deployment model, workflow scope, storage, inference volume, support requirements, and compliance needs. We respond within 2 business days.

Design partners wanted.
Deployed first.
Rendex is selecting a small number of design partner firms across financial and professional services. Design partners help shape workflows, review the architecture and security posture, and receive founding-firm terms locked in for the duration of the partnership.
Direct input on the product

Weekly working sessions with the founding team. The workflows your deal, advisory, and client-service teams actually need become the ones we ship.

Confidential compute available

For firms that require the highest-trust deployment, Rendex offers a Confidential Compute tier with H100-class private inference, hardware-backed attestation, stricter egress controls, and restricted operator access.

First in production

Your firm goes live on infrastructure tuned to the workflows you actually run before anyone else sees it.

Founding terms, locked in

Pricing and seat terms agreed up front and held for the duration of the partnership. No ratchets, no surprise renegotiation.

Become a Design Partner View Trust Center

We respond within 2 business days.

In active development
In development now. Production launch Q2 2027.

We're selecting a small number of design partner firms across financial and professional services to shape the workflows, review the architecture, and deploy first. Founding-firm terms locked in for the duration of the partnership.

Talk About a Design Partnership