Coerentis: The Deterministic Safety Layer for Algorithmic Trading

Think in two layers: log space (the durable record of each Coerentis decision) and the execution path (what actually reaches brokers or your internal OMS).

All three can call the same APIs; the commercial mode describes what you are buying (observe, audit trail, or enforced deployment), not a different “secret” algorithm.

Execution path: strategy → risk/OMS → venue. Anything that can move capital or change exposure.

Log space: the parallel record of Coerentis’s decision for each check: timestamp, inputs you sent, ALLOW or BLOCK, and structured codes (for example drawdown or volatility breach labels). That record is what risk and compliance teams replay, export, or attach to internal tickets.

Shadow prioritises log space without mandating execution changes. Audit hardens log space for production scrutiny. Enforcement ties log space to execution so a BLOCK does not reach the market under the deployed design.

Not automatically from the tier name alone. Licensed — Audit is positioned for teams that need defensible logging, exports, and documentation. Many desks still choose to honour BLOCK in production, but your order defines what you are purchasing.

Enterprise — Enforcement is for organisations that need the execution path wired so BLOCK is operationally binding, with solution engineering and SLAs as agreed.

Always align wiring with your counsel and your regulatory framework.

No. It does not generate alpha, rank names, or fire discretionary trades. Your strategy proposes actions; Coerentis returns deterministic allow/block style outcomes against thresholds you configure.

No. It is infrastructure to constrain tail scenarios and silent failures, not a performance product. Simulations and historical tests are illustrative; they are not promises of future results.

A runaway algorithm keeps trading or sizing in ways that breach risk bounds—often because checks are implicit in strategy code or because failures are silent.

Coerentis centralises gates (drawdown, regime, health-style checks) with explicit reasons. In Enforcement deployments, disallowed trades never reach execution; in Shadow/Audit, you see every would-be block in log space first.

Read how we frame runaway algos in the protection evidence section.

A classic stop-loss often reacts after a position is hurt. A drawdown-style gate evaluates whether the next trade would breach a configured portfolio or strategy drawdown limit before execution, returning a structured block where applicable.

When safety rules live only inside strategy code, they are hard to test, easy to bypass, and painful to explain to regulators. Separating signal generation from risk gates gives you independent tests, cleaner logs, and a single place to prove what blocked a trade.

Coerentis is strategy-agnostic: send proposed trades or context over HTTPS with your API key; receive JSON with allow/block and reasons. Insert between signal generation and routing—see the integration guide for endpoints and a sample middleware pattern.

Instead of opaque “risk says no,” responses include machine-readable codes (for example labels corresponding to drawdown or volatility breaches) suitable for SIEM pipelines, spreadsheets, and audit workpapers—so reviewers see why a trade was blocked without reverse-engineering logs.

Safety decisions come from explicit rules and thresholds you own, not from hidden predictive models. The goal is repeatable, explainable outcomes that stand up in internal and external review.

The stack in the repository is designed to run on your infrastructure (Python/FastAPI). You manage network isolation, TLS, key rotation, and backups. Managed offerings, if any, are defined only in contract.

Treat keys as secrets (environment stores, never front-end). Keep Coerentis close to your execution stack network-wise. Classify logs like any trading data: retention, encryption, and access control are your policies—we provide structured payloads you can map into them.

• Quant and execution engineers building production or shadow stacks
• Risk or technology leadership at funds, props, and systematic desks
• Architects who need deterministic gates and credible log space—not retail signal chasers

Shadow — observe in log space without enforcing on live flow.

Audit — licensed emphasis on forensic logging, exports, and documentation.

Enforcement — enterprise deployment where BLOCK is binding on the execution path as engineered.

Pricing details are on the page above; annual options may apply for Licensed tiers.

For our Audit-first rollout, a limited number of qualified Licensed — Audit customers receive a complimentary extended onboarding and reporting bundle. “25k” is informal shorthand for serious professional scope — not a promised cash amount, not a fixed menu of deliverables. Exactly what you receive is whatever we agree in writing before kickoff. Not a cash rebate.

Indicative criteria (we may decline or waitlist):

• Registered company, fund, or professional trading firm
• Systematic or algorithmic stack with a named technical owner
• Willingness to start in shadow or non-production where appropriate
• Ability to sign an engagement or order (see the scope template; counsel reviews)

Coerentis is patent pending. Website and docs are not legal or patent advice. If you need language for filings, vendor due diligence, or marketing approvals, run final wording past your counsel.