Browse documentation +
OverviewQuickstartInference modesPrivacyLimits & accessDeveloper APIOnchain roadmapErrors
DOCS / PRODUCT / OVERVIEW
SENTIVUM DOCUMENTATION

One interface.
Multiple intelligence layers.

SENTIVUM is a multimodal AI interface for live research, private inference, image creation, and voice generation. This guide documents what is available today and what remains on the roadmap.

PUBLIC BETA

The Studio is live. Developer API access, accounts, payments, and token settlement are not yet public features.

01 / QUICKSTART

Run your first inference

The fastest way to use SENTIVUM is through the web Studio. No account or wallet is required during the current public beta.

  1. 01
    Open the Studio

    Choose the mode that matches your intended outcome.

  2. 02
    Describe the result

    Write a clear prompt with the context, constraints, and format you want returned.

  3. 03
    Run inference

    Research may include sources; image and voice return their native media.

Open SENTIVUM Studio
02 / INFERENCE MODES

Choose the right layer

Each mode is optimized for a different kind of request. Switching modes changes how SENTIVUM routes and returns your prompt.

01● Available

Research

Searches current web sources, synthesizes the findings, and returns source links when available.

BEST FOR
Current events, market context, verification, and source-backed research.
02● Available

Private

Routes text requests through Venice private inference for privacy-sensitive conversations.

BEST FOR
Sensitive drafts, private analysis, and prompts that do not require live web search.
03● Available

Create image

Turns a natural-language description into an original image inside the SENTIVUM interface.

BEST FOR
Concept art, social visuals, campaign directions, and product ideation.
04● Available

Generate voice

Converts written text into playable speech without leaving the Studio.

BEST FOR
Narration drafts, prototypes, accessibility, and spoken content previews.
03 / PRIVACY

Privacy is a route, not a slogan

Private mode uses Venice private inference and is intended for text requests that benefit from a privacy-focused route. Research, image, and voice use separate intelligence services suited to those capabilities.

IMPORTANT

Do not submit passwords, private keys, seed phrases, payment-card data, or information you are not authorized to process. Provider terms still apply.

Local-first history

Studio conversation history is kept in your browser so it can be removed from the device.

Storage-minimized requests

Provider-side storage is disabled where the selected inference service supports it.

Clear routing

The interface identifies when the dedicated Venice private route is selected.

04 / LIMITS & ACCESS

Free within beta capacity

SENTIVUM currently operates as a public beta with shared daily capacity. Requests may be temporarily unavailable when a provider or platform spending limit has been reached.

FeatureAccessBilling
Web StudioAvailableFree beta capacity
Developer APIComing soonNot active
SENT settlementAfter token launchNot active
05 / DEVELOPER API

Build with SENTIVUM

COMING SOON

A developer API with a familiar chat-completions request shape is planned. Public API keys, production endpoints, streaming guarantees, and billing are not available yet.

PLANNED REQUEST SHAPEPreview only
POST /v1/chat/completions
Authorization: Bearer svm_your_key
Content-Type: application/json

{
  "model": "sentivum-core",
  "messages": [
    { "role": "user", "content": "Your prompt" }
  ]
}

This example documents the intended interface. It does not represent an active public endpoint.

06 / ONCHAIN ROADMAP

Robinhood Chain settlement

ROADMAP

After the SENT launch, the planned settlement layer will convert inference usage into credits priced against the token. The product will quote the required amount before an onchain payment is approved.

01RequestUser selects an inference route
02QuoteUsage receives a token-priced quote
03SettlePayment records on Robinhood Chain
ROADMAP

No SENT payment, wallet connection, or onchain settlement is active in the current beta.

07 / ERRORS

Common messages

Daily capacity reached +

The platform or selected route has reached its temporary spending limit. Wait for the limit window to reset and retry later.

Request could not be completed +

The inference service may be unavailable or the request may have timed out. Retry once with a shorter prompt.

No sources returned +

Source links appear only when Research returns usable citations. Rephrase the prompt to request current, verifiable information.