Built to connect.
An API-first architecture designed to sit alongside the custodians, data and systems an institution already runs: custodians and prime brokers, market data, fund administration, order and execution, risk and analytics, and accounting. Open APIs for the rest, with typed SDKs on the roadmap.
Designed to integrate with the venues you already run.
One governed data model underneath execution, market data, custody, administration, compliance and your own warehouse, so the record stays whole however many systems feed it. Most names below are integration targets. The proprietary fund runs today on three of them: brokerage via Alpaca, and market data via Massive and Alpha Vantage.
Execution & brokers
Routes to the venues a mandate already trades. Orders leave only after the kernel permits them.
- Interactive Brokers
- AlpacaIn use · proprietary fund
- Tradier
- Coinbase
- Kraken
- FIX gateways
Market & alternative data
One normalised feed across providers, fed into every observer.
- MassiveIn use · proprietary fund
- Databento
- LSEG / Refinitiv
- Bloomberg
- Alpha VantageIn use · proprietary fund
- ICE Data
Custody & wallets
Connects to wherever the mandate requires assets to be held: bank, broker or qualified custodian.
- BNY
- State Street
- Fireblocks
- Anchorage
- Copper
Fund administration
Reconciles to the administrator of record for NAV strike and close of book. NAV release stays with a named human.
- SS&C
- Citco
- Apex
- NAV Fund Services
Compliance & identity
Screens, verifies and surveils in line with pre-trade checks. Sanctions clearance stays with a named human.
- Chainalysis
- ComplyAdvantage
- Sumsub
- Persona
Data platform & cloud
Streams the insert-only decision ledger into the warehouse you already run.
- Snowflake
- Databricks
- Apache Kafka
- AWS
- Google Cloud
Venues shown are representative of the integration categories Adaptic is designed to support, not a roster of live connections. Three are in use by Adaptic's proprietary fund today: brokerage via Alpaca, and market data via Massive and Alpha Vantage. Every other name is an integration target; specific connectors are provisioned per deployment. Names are referenced for identification only and do not imply a partnership, an endorsement or a client relationship.
The systems your firm already runs.
Organised by counterparty category, not a logo wall. Each category lists the message and data types it carries.
Custodians & prime brokers
- Settlement
- Positions & balances
- Corporate actions
- Securities lending
- Cash & margin
Market data
- Real-time feeds
- Historical feeds
- Reference data
- Corporate actions
- Digital-asset data
Fund administration
- NAV
- Capital accounts
- Statements
- Subscriptions & redemptions
- Fee calculation
Order & execution
- FIX connectivity
- Multi-venue routing
- Allocation
- Transaction-cost data
- Post-trade matching
Risk & analytics
- Factor models
- Stress & scenario
- VaR / CVaR
- Exposure analytics
- External risk systems
Accounting & GL
- Reconciliation
- General-ledger posting
- Portfolio accounting
- Tax lots
- Treasury
On-chain & digital assets
- Custody & wallets
- Settlement corridors
- Tokenised-fund rails
- Stablecoin rails
- Oracles
Identity & productivity
- SSO / SAML
- MFA
- Directory sync
- Document workflows
- Alerting & notifications
Connectors are described as pre-built and planned categories, not a shipped catalogue. Vendor names appear as integration targets, never as a customer roster.
An engineering discipline, not a business-development function.
Each counterparty sits behind a versioned connector with a declared schema, replay semantics, observability and a published runbook. A vendor-side change becomes a typed migration, not a fire drill.
- Typed schemaEvery inbound and outbound message conforms to a strongly typed schema stored with the connector. Vendor schema drift fails at the boundary, never silently.
- Replay & backfillPoint-in-time replay and backfill produce identical canonical events: a framework property, not a maintainer's discipline.
- Observability built inFreshness, throughput, error rate and schema drift are emitted for every connector into one telemetry plane.
- Declared targetsEach connector is designed to declare its feed frequency, maximum lag, replay window and escalation runbook. These are design targets, published with the connector.
- Kernel-gated egressOutbound orders, filings and transfers pass the same kernel that governs trading, so a broken connector cannot emit a message the mandate does not permit. Filings and transfers wait for a named human.HUMAN-GATED
- Change-controlledConnector releases ship with a signed diff, a test run and a release note. Rollback is one command, with version pinning per mandate.
Many connectors, many teams, many backlogs. The platform team patches the graph every time anyone upstream changes a field.
One typed contract maintained as code. A vendor change is a typed migration with a test suite, and every connector event lands in the decision ledger.
An open platform, not a walled garden.
A firm's own team can extend Adaptic without waiting on a roadmap. The schema that drives the platform internally is the one a developer builds against, and every write still goes through the kernel.
- REST, GraphQL and WebSocket APIs: query positions, NAV, risk and the decision ledger; subscribe to real-time events.
- Typed Python, TypeScript and Go libraries generated from the platform schema are on the roadmap. The REST and GraphQL APIs are the interface today.
- Webhooks and streams fan platform events out to a firm's existing workflows.
- A sandbox with synthetic funds and market data, to test integrations without touching production.
Questions an engineering team asks first.
- Can we author our own connectors?That is the design intent. The connector framework is built as a typed SDK, so a firm's platform team will be able to build and maintain connectors against the same contracts Adaptic uses internally.
- How do you handle vendor schema changes?Connectors pin schema versions. A vendor-side change opens a typed migration, tested in the sandbox under change control, rather than breaking the live path.
- Do you support on-chain settlement legs?On the roadmap, under the same connector discipline as every other counterparty: typed, observable, replayable and kernel-gated, for the points where a firm participates on-chain.
- Can we keep data in our own cloud account?The canonical event stream, warehouse and connector artefacts are designed to be deployable into a firm's own AWS, GCP or Azure tenancy where data-residency constraints require it.
See how we plug into your stack.
Share the systems your firm already runs. We map them against the connector categories and show what can be turned on, turned off and consolidated, ahead of a six-week shadow pilot with no capital at risk.