Skip to content

faq.mdx

Latest commit

 

History

History
92 lines (80 loc) · 6.95 KB

faq.mdx

File metadata and controls

92 lines (80 loc) · 6.95 KB
Raw
title FAQ
description Questions we get often.

About SINT Labs

SINT Labs is an AI research and engineering lab. We build three things: the **SINT Protocol** (an open-source governance protocol for autonomous systems), **products** (SINT Console, Avatars, Operators, CMO Operator — unified as SINT OS), and **Studio** work (custom builds and enterprise deployments). No. SINT Labs is an AI infrastructure company. Earlier SINT work (2024–2025) included a Web3 consumer platform with a $SINT token — that work is archived at [/archive](/archive/notice). Current SINT Labs work does not use tokens, does not do on-chain governance, and does not target crypto audiences. **SINT Labs** is the company. **SINT Protocol** is the open-source governance protocol the company publishes. **SINT OS** is the company's runtime that integrates the Protocol with its products (Console, Avatars, Operators, CMO Operator). Think: SINT Labs publishes the Protocol (like Linux Foundation publishes Linux), and SINT Labs also sells products built on it (like Red Hat sells RHEL). SINT Labs is founded and led by Illia Pashkov, based in Los Angeles. SINT Labs is incorporated as PSHKV Inc. in California. As of April 2026 the team is small — founder plus contractors — with hiring underway.

About the Protocol

Because model-level alignment isn't a safety invariant. The same natural-language command issued to four frontier models produces up to a 4.8× spread in out-of-policy attempts. An agent that acts in the physical or economic world needs an authorization layer that's independent of which model is at the wheel. See [Protocol overview](/protocol/overview). Coarser authorization systems (API keys, OAuth, MCP server permissions) have no physics awareness, no tier-based human oversight, no cross-system coordination, and no tamper-evident audit. SINT addresses all four at the protocol boundary. Specifically: SINT tokens carry velocity/force/geofence constraints; SINT escalates to human approval based on a dynamic tier function; SINT's Policy Gateway coordinates across all twelve transport bridges; the SINT Evidence Ledger is SHA-256 hash-chained. Current version is v0.2. The reference implementation has 1,772 tests passing, the 157-test conformance suite is green, and components are used in production on internal SINT Labs deployments and with pilot Studio clients. It's not at v1.0 yet, which means breaking changes can still happen in minor versions. Review the [changelog](/changelog) before deploying. Code is **Apache-2.0**. Specification is **CC BY 4.0**. See [licenses](/legal/licenses). Yes. The reference implementation has no external dependencies at runtime. Postgres and Redis are the only required infrastructure; both can run on-prem. TEE attestation is optional and can be configured against Intel SGX, AMD SEV-SNP, or AWS Nitro Enclaves. Yes. SINT is model-agnostic. The CSML score measures each model's in-situ behavior and adjusts the tier escalation function accordingly. Any LLM-driven agent can be governed by SINT, including open-weights models like Llama and Mistral.

About the products

**SINT OS** is the runtime — the integration layer that connects Protocol + products. It's headless. **SINT Console** is the visual control surface on top of OS. Think of OS as the kernel and Console as the desktop environment. Technically yes — the Avatars repo runs standalone. Practically, the avatar is at its best when wired into OS because the Conversation Compiler and widget system use state from the Protocol's Evidence Ledger and Policy Gateway. No. Sintra AI is an unrelated product with a similar name. CMO Operator is SINT Labs' content production pipeline, built on SINT OS and governed by SINT Protocol. We don't build or integrate with Sintra AI.

About Studio

SINT Studio is the agency/dev-shop arm of SINT Labs. It's one of three divisions (alongside Protocol and Products). We build agentic systems for companies who need the outcome without building the stack themselves. See [sint.gg/studio](https://sint.gg/studio). Starter engagements start at $25K for 2–4 weeks (one agentic loop, in production). Setup engagements start at $75K for 6–10 weeks (full operator deployment). Retainer engagements are monthly and priced per scope. See [sint.gg/studio](https://sint.gg/studio). Services firms $3M–$20M revenue, 15–80 employees — professional services, legal, logistics, healthcare admin, real estate, marketing agencies. Companies with repeatable operational workflows where autonomous agents close quantifiable losses.

About archived work

The $SINT token was part of SINT Hive, our earlier consumer Web3 platform. In April 2026 SINT Labs refocused on physical-AI governance. The Web3 consumer work is paused, not sunset. Existing commitments stand. See [/archive/web3-platform](/archive/web3-platform). Still available. Community support continues. See [/archive/web3-platform](/archive/web3-platform). Those announcements are historical. See [/archive/partnerships](/archive/partnerships). They don't describe current SINT Labs work.

Still have questions?