Agentic Connectors

Connectors that build themselves. Show HyperTrail a sample event, describe the mapping in plain language, and the connector generates, deploys, and runs itself: inbound events onto your digital twins, outbound actions into your systems.

9 connector types · 16 prebuilt integrations

The reference

Nine connector types.

Every connection between your systems, your twins, and your agents is a first-class primitive. Warm means you own it: your events, your keys. Cool means HyperTrail generates and operates it for you.

Core

The five connections every Trail uses.

Events In

S2EYours

Your business events land on digital twins through AI-assisted mapping and production-ready REST endpoints.

Context to Agent

E2APlatform

Agents receive real-time entity change events with full entity context. Filtering rules included.

Outcomes Recorded

A2EPlatform

Maps agent actions to entity operations, writing decisions and outcomes back to the twin.

Actions Out

A2SYours

Agent decisions become structured requests against your downstream APIs, through your keys.

Facts Shared

E2EPlatform

Real-time replication between entities. One upstream event can update thousands of downstream twins.

Power

For multi-agent and outbound-data architectures.

Twin Export

E2SYours

Automatically propagates entity changes out to your systems: warehouses, CDPs, BI dashboards.

Direct Trigger

S2APlatform

A system event invokes an agent directly, no twin threshold in between.

Agent Handoff

A2AComing SoonPlatform

One agent triggers another with full context. Multi-agent orchestration as a primitive.

Specialized

For multi-modal experiences.

Vision Enrichment

P2EPlatform

Turns images into structured facts on the twin, so agents process them alongside entity data.

The lifecycle

Loop until it works.
Publish when it does.

Every connector goes through a build, deploy, test loop before anything touches production. Publish locks the version, and you get a live endpoint and API key the same moment.

Build

Give it a sample payload and a sentence. An agent generates the mapping, the transforms, and the identity stitching.

Deploy

One click spins up isolated serverless infrastructure for the connector.

Test

Send sample events. Inspect the facts that land on the twin. Revise and redeploy until it behaves exactly right.

Publish

When it works, you publish. The version locks and the connector is yours.

Send events

You get an endpoint and an API key on the spot. Point your system at it and events flow immediately.

hotel_crs_events · Events Inbuild #3

Build · mapping generated

sample payload + "map check-ins to a Stay fact, identify by loyalty number"

Deploy · connector live in sandbox

isolated serverless runtime · 1.2s

Test · sample event sent ✗

Stay fact landed on guest #4821 · room name wrong

Build · mapping revised

RmType code table corrected · redeployed

Test · sample event sent ✓

Stay fact correct · timezone verified · identity stitched

Publish · version locked

v1.0.0 · published by you

live · ready for events

POST https://events.hypertrail.ai/v1/hotel-crs

x-api-key: ht_live_••••••••7f2a

Inside the build step

Any source event. One sentence.
The right fact.

No mapping UI, no transformation scripts. An agent reads your sample payload, reads your sentence, and generates the mapping: field extraction, transforms, and identity stitching included.

Source eventany system, any shape

{

"EvtType": "RES_CHECKIN",

"GuestNum": "88213",

"LoyaltyNo": "GLD-4821",

"RmType": "DLX-K",

"TsEpoch": 1784206931,

"PropCode": "LV-03",

}

a real payload from your Hotel CRS · vendor field names and all

Your instructionplain English

"Map check-ins to a Stay fact on the guest profile. Identify the guest by loyalty number."

Generated mappingagent-built

LoyaltyNo → guest identity

stitch key · matches loyalty_id on the twin

EvtType → fact type: Stay

"RES_CHECKIN" recognized as a check-in

TsEpoch → check_in_at

epoch → ISO 8601, property timezone applied

RmType → room: deluxe king

vendor code table inferred from samples

PropCode → property: Las Vegas 03

resolved against your property list

✓ connector generateddeployed serverless · no code written

The same capability runs in both directions: System-to-Entity (S2E) connectors map source events into facts on your twins, and Agent-to-System (A2S) connectors map agent decisions into structured requests against any downstream API. The catalog below covers what agents can act through today.

The Actions Out catalog

Where your agents can act today.

Prebuilt Agent-to-System (A2S) connectors. Agents use them to push decisions, synchronize systems, and send personalized content through the tools your teams already use.

Marketing Automation Customer Communications CRM & Sales Customer Data Platforms Advertising & Paid Media Push & In-App Messaging Support & Customer Experience Productivity & Collaboration

Marketing Automation

Trigger personalized campaigns, journeys, and messaging in real time from AI agent outputs.

Braze

Push real-time user attributes, trigger campaigns, and update audience segments with HyperTrail agents.

Iterable

Sync user profiles and trigger cross-channel workflows based on AI-driven behavioral patterns and predictions.

Customer Communications

Reach customers across SMS, email, voice, and chat with AI-orchestrated messaging.

Twilio

Send AI-personalized SMS, MMS, WhatsApp messages, and voice calls triggered by agent decisions.

SendGrid

Deliver transactional and marketing emails with dynamic content generated by HyperTrail agents.

CRM & Sales

Keep CRM records enriched and up to date with AI-generated insights and next-best actions.

Salesforce

Update leads, contacts, and opportunities with AI-scored attributes and recommended actions.

HubSpot

Enrich contact properties and trigger workflows based on predictive agent outputs.

Customer Data Platforms

Feed AI-enriched profiles and events back into your CDP for unified customer intelligence.

Segment

Emit enriched events and traits from HyperTrail agents into Segment for downstream activation.

mParticle

Sync AI-computed audiences and user attributes into mParticle for cross-platform orchestration.

Advertising & Paid Media

Activate AI-built audiences and suppression lists across ad platforms for smarter spend.

Google Ads

Push Customer Match lists and conversion events driven by AI agent predictions.

Meta Ads

Sync Custom Audiences and offline conversions from AI-scored customer segments.

Push & In-App Messaging

Deliver contextual push notifications and in-app experiences powered by real-time AI decisions.

OneSignal

Trigger personalized push notifications and in-app messages based on AI-predicted user intent.

Airship

Orchestrate mobile engagement with AI-driven message timing, content, and channel selection.

Support & Customer Experience

Route, prioritize, and enrich support interactions with AI-generated context and recommendations.

Zendesk

Auto-tag, prioritize, and enrich tickets with AI-generated customer context and suggested resolutions.

Intercom

Surface AI-driven customer insights and next-best actions directly in live chat and inbox.

Productivity & Collaboration

Connect AI agents to workplace tools for automated document management, scheduling, and communication.

Google Workspace

Coming Soon

Read emails, search documents, and manage calendar events through Google's MCP-native agent interface.

Microsoft 365