Skip to content
Brevor.
B2B

BridgeAPI

Bridge API is the same routing engine that powers Bridge Workspace, exposed as a programmable endpoint. One call between your application and any model provider. SLA-backed, audit-complete, in production at 1.8 billion routing decisions per quarter.

1.8B routing decisions last quarter

Talk to sales →Documentation
WHY THIS EXISTS

Some teams want the layer without the workspace around it.

Not every Bridge customer wants a control plane, an audit UI, and audience profiles. Some teams have all of that already. What they want is the routing engine — exposed cleanly, with the audit trail intact, with the SLA written down.

Bridge API is that surface. Same engine as Workspace. Same audit primitives. Same compliance posture. Different shape: one endpoint, one schema, one billing line.

The Affinity Federal Credit Union deployment is documented as a Bridge API case study. Ninety days in production, zero unsanctioned outputs, regulator-accepted audit trail at day 60.

HOW IT WORKS

Three steps from prompt to decision.

Single endpoint

Your application calls one Bridge API endpoint. Bridge handles model selection, context loading, and behavioral policy. Your application receives the routed response and the audit reference.

01

Per-call policy

Audit policy, context profile, and escalation rules pass per call. Defaults are configured per API key. The policy is enforced at routing, not by your application.

02

Audit reference returned

Every call returns an audit reference. Your application stores the reference. The full audit record lives in Bridge's audit storage and is queryable via the audit API.

03
BUILT FOR

Where Bridge API lives in production.

Product engineering teams

Teams that want the routing primitive without the surrounding workspace. They have their own application surface and their own access control.

Used by a fintech's risk-modeling team routing model calls from their proprietary application.

API-native deployments

Customers whose AI usage is programmatic and doesn't pass through a UI Brevor would build for them.

Used by a logistics company's dispatch system routing model calls 200K times daily.

Multi-product enterprises

Customers using Bridge API behind multiple applications in their organization, each calling the same Bridge deployment.

Used by an insurance carrier with API integration across claims, underwriting, and member-services systems.

Existing-stack augmentation

Customers who already have model governance, model serving, and policy documentation, and want Bridge as the runtime enforcement layer.

Used by a regional bank already running an AI policy framework and adding Bridge as the enforcement primitive.

SPECIFICATIONS

What ships in every deployment.

Deployment
SaaS, single-tenant, or air-gapped
Median latency (p95)
78ms behavioral routing overhead
Throughput
Tested to 28,000 RPS per API key
Concurrent contexts
64 context modules loadable per request
Compliance
SOC 2 Type II, HIPAA, GDPR, ISO 27001, FedRAMP Moderate
Audit trail
Full decision provenance, replayable from inputs, audit reference returned per call
SDKs
Python, TypeScript, Go, Java, Ruby
SLA
99.95% routing availability, 24/7 incident response, error budget published quarterly
INTEGRATION

One endpoint between your application and any model.

curl
curl https://api.brevortech.com/v1/route \
  -H "Authorization: Bearer $BREVOR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Generate a compliance summary for transaction #92447.",
    "context_profile": "naic_finserv_v3",
    "audit_policy": "ny_dfs_strict"
  }'

Full API reference at /docs →

IN PRODUCTION AT
Affinity Federal Credit UnionNavy Federal Credit UnionStifel FinancialAmeriprise FinancialWintrust FinancialRenaissanceReHub GroupKratos Defense & Security Solutions
FREQUENTLY ASKED

Questions buyers ask before scoping.

Bridge API is the routing primitive. Workspace is the full enterprise deployment built on Bridge API. Teams who already have a control plane and audit UI typically pick API.

Median p95 is 78ms for the routing decision plus context load. Add the model provider's latency on top. Latency budgets are documented in the latency-budgets research post.

Every API response includes an `audit_id`. Your application stores the ID. The audit record stays in Bridge's storage and is queryable via the audit API, with default 7-year retention.

Yes, single-tenant and air-gapped deployments are supported. The same API contract holds. SLAs differ based on infrastructure ownership.

All major providers via Bridge's routing layer. The provider list is per-deployment configurable. Adding a provider is a deployment configuration change, not a code change.

Per deployment, with throughput tiers. Sales scopes on a call based on expected RPS, audit retention, and SLA tier.

See Bridge API in your environment.

Talk to sales

Backed by Kyros Holding Inc.