Thyn

Intelligence.
At speed.

On-device AI engines engineered for performance.

1000ms

Down to nanoseconds.
By design.

Selected runtime workloads execute in nanoseconds. Local-first. Developer controlled.

AI shouldn’t need a round trip to the cloud.

Intelligence should run closer to users, closer to data, and closer to the decisions it shapes.

Now Live

Algenta

Engines for agents, workflows, memory, simulations, and governed execution.

Sense operational state → plan → execute workflow → evaluate outcome.

Visit algenta.ai
01 / 07
Coming Soon

Decimas

Systems for execution strategies, market simulations, risk, and quantitative decisioning.

Observe market → simulate outcomes → route execution → measure risk.

Get updates
02 / 07
Coming Soon

Escrypto

Infrastructure for verification, replay, digital assets, and self-hosted financial systems.

Authorize action → verify state → execute → prove history.

Get updates
03 / 07
Now Live

Codna

Maps your entire repo in milliseconds — for zero tokens — then sends agents in with exactly the context they need.

Map repository → understand structure → retrieve exact context → guide the agent.

Visit codna.ai
04 / 07
Now Live

Telys

On-device memory engine for AI agents — millisecond recall from a single contiguous scan, with zero cloud round-trips.

Ingest context → index on-device → recall in a single scan → ground the agent.

Visit telys.ai
05 / 07
Now Live

Cohenta

Autonomous growth systems for campaigns, attribution, experiments, and content intelligence.

Generate hypothesis → launch experiment → attribute signal → optimize spend.

Visit cohenta.ai
06 / 07
Now Live

SQAI

The governed, read-only structured-data tool for AI agents — typed plans, deterministic execution, provenance hashes on every answer.

Receive question → author typed plan → check contract & policy → execute deterministically → stamp provenance.

Visit sqai.com
07 / 07

Built for developers. Faster than the network.

Thyn is the parent company behind specialized engines for AI, trading, cryptography, code intelligence, memory, and autonomous growth.

Latest.

June 2026

Telys joins the portfolio

An on-device memory engine for AI agents — millisecond recall from a single contiguous scan, with zero cloud round-trips.

April 2026

Decimas enters development

Trading intelligence for simulations, execution strategies, and quantitative systems — built to observe markets, model outcomes, and route execution under real latency budgets.

December 2025

Algenta goes native

Algenta’s selected execution paths move onto a native, accelerated runtime — bringing more of the agent loop on-device for lower latency and tighter control.

Bring intelligence closer.