AGENT-NATIVE · THREE DOORS · GRADED IN PUBLIC

QL Agent Data Plane

The private market, as a data plane for AI agents. One coverage universe — every US stock and ETF plus 2,300+ private companies and 350+ VC/PE firms — behind three doors an agent can actually walk through: a one-click MCP connector, a self-serve REST API, and a paid endpoint that needs no account at all. No seat license. No sales call. Every fact dated, sourced, and graded in public.

Connect your AI in one click REST API docs llms.txt

Three doors, one plane

Door 1 · MCP

One-click connector

180+ tools over OAuth — signals, fundamentals, options flow, private-market valuations, investor graphs, IPO research. No API key to provision; works with Claude, ChatGPT, Cursor, and any MCP client.

https://quantlogix.ai/api/mcp
Door 2 · REST

Self-serve API + SDK + CLI

Versioned /api/v1/* with self-serve ql_ keys, a zero-dependency JS SDK, and the ql CLI. Provision a key from your profile — no contract, no procurement cycle.

curl -H "Authorization: Bearer ql_..." \
  https://quantlogix.ai/api/v1/ask -d q=...
Door 3 · No account

Pay-per-answer

One grounded answer for $0.50, settled in-band over the open Machine Payments Protocol. The payment is the authentication — nothing to sign up for. Agents can also refill their owner’s AI credits the same way.

POST /api/v1/agent-pay/answer?q=...
→ HTTP 402 → pay → answer + receipt

Facts with receipts, not just facts

Data platforms assert; this one shows its work. Every valuation carries its date and source. Paid answers ship a SHA-256-stamped answer_record. Directional stances are graded against the tape at T+30 and published — hits and misses with equal billing. The private-market price series is graded forward-only against the next priced round, against the only baseline that matters: would you have done better just quoting the last round?

Paid answers served


every one hash-stamped

Stances graded


against real prices at T+30

The graded record

QL Ledger — every call SHA-256 chained · Private Price — the graded valuation series · Signal Truth — verified receipts

Freshness you can check, not take on faith

Every 6 hours

The private-market deep-enrichment pass sweeps the roster — valuations, leadership, secondary prints — with per-field provenance.

Daily

The private-market daily pipeline turns filings, rounds, and tenders into dated, sourced events. Public-market signals recompute on market data continuously.

Dated on every read

Every mark carries its as-of date and anchor, so staleness is visible instead of hidden — the honest alternative to quietly serving last year’s round.

The honest comparison

Seat-licensed incumbents like PitchBook are excellent databases sold to humans. This plane is built for the agent as the first-class consumer — and the comparison below concedes the row they win.

Seat-licensed incumbentsQL Agent Data Plane
Access modelAnnual per-seat contracts, sales-mediated; list pricing publicly reported in the tens of thousands per seat per yearSelf-serve — free tier, Pro subscription, or $0.50 per answer with no account at all
AI / MCP accessMCP connector gated on holding a licensed seatOne-click OAuth for any MCP client — no seat, no sales call
APIEnterprise agreements; no self-serve keySelf-serve REST + SDK + CLI, provisioned from your profile
Accuracy recordNot publishedForward-graded in public — hits and misses with equal billing, hash-chained
Payments an agent can makeNone — a human signs the contractHTTP-402 over the open Machine Payments Protocol; the agent pays in-band
BreadthMillions of companies — if you need the long tail, they win2,300+ private companies + every US stock and ETF — the names agents actually ask about, re-enriched every 6 hours

Incumbent characteristics are based on publicly available information as of August 2026 and may change — verify with the vendor. QuantLogix deliberately does not compete on long-tail breadth.

Everything in the plane

MCP connector docs One-click setup for Claude, ChatGPT, Cursor, and any MCP client REST API docs Versioned /api/v1 endpoints, tiers, and OpenAPI spec QL Agent Answers Pay-per-answer — no account; the payment is the authentication QL Datasets Bulk, citable datasets with stable URLs Agent API & JS SDK Run scheduled AI analysts over the same data plane Agent Studio Build an agent in natural language ql CLI Zero-dependency terminal client over the MCP door — install via /cli/ql.mjs llms.txt The machine-readable index — built for agents, not crawlers

Market intelligence, not investment advice. MCP and REST access are metered on the account’s plan; pay-per-answer settles via Stripe under the Machine Payments Protocol. The accuracy record and its grading methodology are public — start at the QL Ledger.