What is VoltageGPU Confidential Chat?

A ChatGPT-style interface running inside Intel TDX hardware enclaves. Prompts and responses are encrypted in CPU memory at runtime and never leave the sealed Trust Domain. VoltageGPU cannot read user data, cannot log it, cannot train on it.

Do I need an account to try it?

No. Anonymous visitors get 3 free messages on Qwen3-32B-TEE without any signup, no credit card required. After the free trial, signup unlocks 50 free messages per month, Plus $20/mo unlocks 2000 messages with Qwen3-235B 262K context.

How is it different from ChatGPT Plus?

Same $20/mo price point but sealed in Intel TDX hardware. OpenAI trains on user data by default. VoltageGPU Plus cannot technically read user data — memory is encrypted with CPU-fused keys and even the hypervisor is outside the trust boundary. EU company (France, SIREN 943 808 824), native GDPR Article 28 DPA available.

Which TEE models are available?

3 models sealed in Intel TDX: Qwen3-32B-TEE (40K context, fast), Qwen3-235B-A22B-Instruct-2507-TEE (262K context, flagship, Plus/Pro), DeepSeek-R1-0528-TEE (163K context, reasoning model, Enterprise).

Can I use it for regulated work (legal, medical, finance)?

Yes. Beyond the generic chat, 8 specialized agents are available: Contract Analyst, Financial Analyst, Medical Records, Compliance Officer, Due Diligence, Cybersecurity, HR, Tax. GDPR Article 28, HIPAA, DORA, NIS2 ready. DPA available on request.

Yes — api.voltagegpu.com/v1 is OpenAI-compatible, pay-per-token starting at $0.15/M input tokens on Qwen3-32B-TEE. Drop-in replacement for the OpenAI SDK (change base_url).

How is the FP&A agent different from Pigment AI or Vena copilot?

Pigment AI and Vena copilot are AI features inside their respective FP&A platforms. The VoltageGPU FP&A agent is a confidential layer that works on top of the financial data export the CFO already has — Excel, Anaplan, Pigment, Vena, Workday Adaptive, Oracle EPM, SAP Group Reporting. The differentiator is Intel TDX hardware sealing for management-pack confidentiality.

Why does management-pack confidentiality matter?

Management accounts, monthly variance, forecast revisions and board memos contain non-public information. Pasting that material into ChatGPT Enterprise or Microsoft Copilot creates a sub-processor relationship most CFOs do not want logged against IPO-stage, M&A-sensitive or competitively-sensitive data. TDX sealing makes the inference loop provider-blind.

Financial Compliance / FP&A TDX Attestation Live

For CFOs, Heads of FP&A and finance directors

FP&A AI Agent — Confidential Financial Planning.

Drafts variance analyses, forecasts, board memos and sensitivity tables on the management pack the CFO refuses to paste into ChatGPT. Sealed in Intel TDX. EU-hosted.

Works on top of Excel, Anaplan, Pigment, Vena, Workday Adaptive, Oracle EPM and SAP Group Reporting exports. The agent drafts; the CFO and finance team sign.

See pricing Back to financial compliance hub

Variance analysis

Month-end variance, decomposed and explained.

The agent ingests actuals vs budget and actuals vs prior period from the management pack export. It decomposes the variance across the standard FP&A drivers — price, volume, mix, FX translation, FX transactional, one-offs — and drafts an explanation paragraph in the CFO's preferred tone.

Typical use: the FP&A analyst pastes the export, gets a draft variance commentary in minutes instead of hours, then enriches with the operational context only the business knows. The CFO signs the final commentary.

Forecast drafting

Top-down and bottom-up forecast drafts.

For top-down forecasts, the agent uses the historical seasonality and the latest actuals to draft a monthly trajectory for the rest of the year, with explicit assumptions on growth rate, margin trajectory and run-rate cost. For bottom-up forecasts, the agent helps consolidate department submissions into a coherent narrative around drivers.

The agent does not autonomously commit a forecast to the FP&A platform. It drafts; the FP&A team enters or accepts the values inside their existing tool (Pigment, Vena, Anaplan, Workday Adaptive). This keeps versioning, approval flow and audit trail inside the system of record.

Board memo drafting

Board memos, drafted in the CFO's voice.

Board memos and audit committee memos pull from the same management pack but need a different tone — concise, prioritised, decision-oriented, with the right level of granularity for non-executive directors. The agent drafts the memo from the pack and the CFO's prior memos, learning the tone over time.

Because board memos can contain inside information, IPO-stage material or M&A sensitivity, hardware-sealed inference is the table-stakes feature. Pasting board-pack drafts into ChatGPT Enterprise creates a sub-processor relationship most legal / compliance teams will not approve. TDX sealing closes that gap.

Sensitivity-table generation

Sensitivity tables, across the drivers that actually move the P&L.

For each forecast scenario, the agent drafts a sensitivity table across the relevant drivers — pricing, volume, FX (EUR/USD, EUR/GBP, etc.), interest rates, churn or retention, COGS inputs. The output is a formatted table the FP&A team can paste into the management pack.

The agent labels assumptions explicitly and explains the second-order effects (a price-volume trade-off, a pricing-pass-through assumption on inflation, etc.). The CFO and FP&A lead validate the assumptions before the table goes into the deck.

Capabilities

What the agent drafts.

Variance commentary

Decomposed across price, volume, mix, FX translation/transactional, one-offs.

Forecast drafts

Top-down and bottom-up trajectories with explicit driver assumptions.

Board memos

Decision-oriented memos in the CFO's voice, learned from prior memos over time.

Sensitivity tables

Across the drivers that actually move the P&L — labelled assumptions, second-order effects.

Driver-tree narratives

Connects KPIs to financial drivers in a narrative format suitable for the management pack.

Confidential by hardware

Inference inside Intel TDX. Provider-blind. EU-only data path. ECDSA attestation per request.

Comparison

VoltageGPU FP&A agent vs Pigment AI vs Vena Copilot.

Pigment AI and Vena Copilot live inside their respective FP&A platforms. VoltageGPU is a confidential layer that works on whatever export the CFO already has. Different access surface, different confidentiality posture.

DimensionPigment AIVena CopilotVoltageGPU

Where it livesInside Pigment platformInside Vena platformOn top of any export

HostingPigment cloudVena cloud (Microsoft Azure)EU TDX (VOLTAGE EI, FR)

Confidential computeStandard cloudStandard cloudIntel TDX hardware enclave

Per-request attestationNoNoECDSA quote per request

Variance commentaryYesYesYes — TDX-sealed

Board memo draftingLimitedLimitedYes, in CFO voice over time

Works on Excel / SAP / Oracle EPMPigment ecosystem onlyVena ecosystem mostlyYes, any export

Pricing

Same tiers as the financial compliance pillar.

Starter $349/mo (small FP&A team), Pro $1,199/mo (10 seats, full Article 30 pack), Enterprise $3,499/mo (dedicated tenant, DeepSeek-R1-TEE for board-pack reasoning, SOC 2 / ISO 27001 path).

See full pricing

FAQ

Questions CFOs ask first.

Is the agent producing investment guidance or capital-markets advice?

No. The agent drafts internal management content. It is not financial, investment or capital-markets advice. CFO-level conclusions, forecasts and board recommendations remain the responsibility of the human finance team.

Can it write directly into Pigment / Anaplan / Vena?

Not by default. The agent drafts; the FP&A team enters or accepts the values inside their existing system of record. This keeps versioning, approval flow and audit trail where they belong. Custom write-back integrations are scoped on the Enterprise plan.

What about IPO-stage or M&A-sensitive data?

That is the canonical use case. Hardware-sealed inference is precisely the differentiator most legal / compliance teams require before they will let board memo drafts touch any AI tool. The TDX attestation log gives general counsel the evidence trail.

Does it work in multiple languages?

Yes. The agent runs on Qwen3-235B-TEE and DeepSeek-R1-TEE — strong in English, French, German, Spanish, Italian and most major European languages. The board-memo voice is learned per CFO, in their working language.

Financial compliance pillar

Back to the AML / audit / FP&A hub for finance teams.

AML Monitoring AI

Suspicious transaction triage. FATF / 6AMLD aligned.

Audit Automation AI

Walkthroughs, sample-test summaries, SOX 404 narratives.

Sovereign Agentic AI

The hub: vertical agents for legal, finance and regulated professionals.

DORA AI compliance

How VoltageGPU lines up with DORA Article 28-30.

EU AI Act compliance

Article 26 deployer evidence, model cards, logging.