Trading infrastructure builds
Venue scanners, resolution matchers, execution rails, risk systems, and honest dashboards for your strategy. Most builds land between $10,000 and $30,000 and ship in three to six weeks, scoped in writing before a line of code exists.
- Full source repository, owned 100% by you
- The written spec you signed off, as documentation
- Complete test suite and the adversarial review record
- Deploy rails, monitoring, and an operations runbook
- 30 days of post-handover fixes included
Strategy risk audit
For bots that fill fine but still leak money. You send pairs and fills, we run them through the failure catalog behind Basis: mismatched resolution rules, bait depth, fee-blind math, noise-floor fills. You get a written report naming every leak found, the estimated cost of each, and the fix, then a walkthrough call.
- NDA first, if you want one, before any data moves
- Report and call within 5 business days of data received
- Your data serves your audit only, deleted after
If the report finds no concrete leak, you pay nothing.
Not sure yet? Email three fills you don't understand for a free read, no charge.
The trap map
A feed of cross-venue pairs that look identical and are not, flagged and scored by the matcher. 308 pairs tracked today, zero rated safe to trade. In development.
Join the waitlist
The codeFull source, spec, tests, and docs are yours. Nothing of yours is kept or resold.
The continuityEvery deliverable is built to outlive the vendor: written spec, full tests, runbook. No bus factor buried in one head.
The support30 days of fixes included on builds. Ongoing retainer optional, never required.
The paperNDA available before any data or strategy detail moves. Engagement terms in writing.
BriefOne call. What you trade, where it breaks, what the machine must never do.
SpecA written plan with gates and rails. You sign off before a line of code exists.
BuildImplementation runs against the spec, test-first, in tight review loops.
AttackAdversarial review tries to break every merge before it ships.
VerifyDeployed against the real venue and watched live before handover.
Audits skip the build steps: data in, catalog run, written report and call inside five business days. And the process is not trading-specific: the same spec-first discipline ships data pipelines, internal dashboards, and automation for operations teams.
Do you trade money for clients?
No. Basis does not accept or manage outside capital, and we do not sell signals. We build and audit the machinery. Your capital, your keys, your account.
Who owns what I pay for?
You do, fully: source code, the written spec, the test suite, the docs. We keep nothing of yours and resell nothing of yours. Your strategy details never feed anyone else's build.
What happens after handover, or if you vanish?
Builds include 30 days of fixes, and every deliverable ships with the spec, full tests, and a runbook so any competent engineer can run and extend it without us. The bus factor is designed out of the deliverable, not promised away. A retainer is available if you want us on call.
Is my strategy confidential?
Yes. NDA before any data moves if you want one. Audit data is used only for your audit and deleted when it closes. Nothing you share is traded on, benchmarked against, or reused.
Have you delivered for clients before?
The operator has shipped production systems since 2022, including media pipelines and automation for brands like Lodha, Max Estates, and Hero Homes, and runs the SimplifyGenAI studio. Basis is the public, checkable proof of the trading-infrastructure process. Engagements built on it are new, so founding clients get founding pricing, direct operator access, and become the reference cases.
Why should the refusals convince me?
Anyone can show a green screenshot. A machine that journals every decision, including a 10.95% edge it declined with a written reason, is much harder to fake. The discipline is the product, and this site runs on the same records.
How does a two-week build carry 869 tests?
Because nothing starts without a written spec, nothing merges without adversarial review, and everything mechanical in the toolchain is automated: scaffolding, test harnesses, refactors. The judgment stays human and accountable. And at runtime, nothing generative sits anywhere near the trade loop: the machine that touches markets is pure, testable arithmetic.
What stacks do you work with?
Python and TypeScript by default. Venue APIs, on-chain execution, market data pipelines, SQLite or Postgres ledgers, systemd or containerized deploys. If it is a market and it has an API, it is in scope.