The information layer for autonomous trading and decision-making.
Marking lets software buy fresh decision-ready information without an account—and verify it before acting.
Autonomous software can access data. It cannot easily assemble trustworthy information.
An agent needs the information required to decide.
Is there an executable BTC arbitrage opportunity right now?
What is the current BTC market-making state?
Fair-value inputs · imbalance · volatility · liquidity · funding
Did this event happen?
Sources · evidence · resolution policy · attestation
Marking returns trusted information. The agent owns the decision and execution.
Marking turns fragmented data into decision-ready information.
Why now: Software can increasingly act and spend, but assembling trustworthy information still requires accounts, credentials, several APIs and custom analytics.
The same information layer can power many autonomous strategies.
Available today: canonical crypto prices and one event-resolution prototype. Strategy-specific intelligence products are next—not claimed as complete.
One testnet payment unlocked a signed live stream.
Certification report evidence. Base Sepolia testnet. Three latency samples—not a percentile or geographic SLA. Payment and access are implemented; decision and execution remain with the consumer.
Trust travels with the information.
The recipient verifies the object—not a screenshot, dashboard or database assertion.
Representative implemented schema fields. Publisher-origin signatures and broader intelligence envelopes remain planned.
The information path stays fast; payment and durable state stay separate.
200 feeds · 1,550 obs/s · 0.110526 ms P95 source-received → canonical-signed
60-second scoped local Rust benchmark; 93,001 observations. Not production, geographic or consumer-delivery certification. Rust remains shadow-only.
Marking sits above fragmented information sources.
A new infrastructure problem created by autonomous software
Marking starts with financial information because it is frequent, valuable and measurable; the same trust model extends to final outcomes.
provide valuable raw and canonical data
assembles decision-ready information for autonomous consumers
discover → understand → pay → consume → verify → decide
Three product types match three information jobs.
Raw and canonical information
BTC/USD for 30 days = $51.84Market state, volatility, liquidity, funding and risk
Initial pricing hypothesisA specific answer: opportunity, funding, liquidation or outcome
Initial pricing hypothesisThe technical preview is pre-revenue. An invited external-user cohort will measure repeated usage and willingness to pay. Provider rights and fees remain release gates; monthly plans may coexist with machine-native usage pricing.
Every new consumer can make the information network more valuable.
Find a narrow cohort of developers who repeatedly pay for and verify decision-ready information.
FOCUS → Reach weekly retained external machine consumers before expanding supply.
The core information infrastructure exists. The product layer is expanding.
Make information programmable for autonomous software.
Technical execution: deployed preview + certified payment flow + scoped 200-feed local benchmark.
14+ years across JPMorgan Chase and Amazon; former VP Technical Program Manager in Digital Markets Execution & Technology; 4+ years in blockchain.
About 14 years across JPMorgan Chase, Amazon and crypto GTM; institutional-client, ETF-marketing and customer-acquisition experience.
FOUNDING TEAM → PRODUCT & TECHNOLOGY + GTM & OPERATIONS