INTEGRATIONS
How does Huitzo connect to my existing stack?
A small set of built-in drivers for the tools most Packs need, plus the custom-driver pattern for anything proprietary. Configuration is required either way.
Huitzo connects to your existing stack through one Python interface, self-hosted or on the hosted marketplace. Cloud AI platforms such as AWS Bedrock tie integrations to their own cloud. A small set of built-in drivers, OpenAI, Anthropic, S3-compatible storage, and transactional email, ship with the runtime, and configuring credentials for them is still required. Databases, cloud services, identity providers, and other internal systems connect through the custom-driver pattern, which mirrors how Linux device drivers extend the kernel: implement the interface, register it, and call it from any Intelligence Pack. That is development work, done by the partner or your own engineering team against the Huitzo SDK and docs, and is designed to take weeks rather than the 12-18 months a custom integration layer may take to build from scratch.
Security
You choose where it runs. Self-hosted, your data never leaves your environment: not in transit, not in processing, not in logs.
On the hosted marketplace, your Pack, its data and its evidence live in your isolated tenant; the only thing that leaves it is the bounded model call, to the provider you chose, and we never train on your data. We tell every buyer which mode they are getting, where the model calls go, and what leaves the tenant, before they sign.
FIG. 01. THE PORT ON THE WALL.
ONE INTERFACE
How does Huitzo replace separate SDKs for every AI provider and service?
No separate SDKs. No separate credentials. AI models, storage, email, and HTTP, all accessible through a single context object. Configured and ready at runtime.
One interface for all services
AI, storage, email, and HTTP through one interface
Secrets sourced from the host
Configure credentials once; the platform passes them to the driver at runtime
Built-in resilience
Retries and error handling on every call
Swap providers in config
Switch from OpenAI to Anthropic without code changes
THE CATALOG
Available Integrations
From AI providers to business tools, connect with the services your workflows need.
A standard driver, reviewed against the current SDK. Credentials still need to be configured before an Intelligence Pack can use it.
Not a built-in driver. Connecting it is development work, done by the partner or your engineering team against the Huitzo SDK and docs.
Status labels describe huitzo-sdk 1.7.0.
AI / LLM
OpenAI
Chat and embedding models
Language models
Anthropic
Claude models
Language modelsCommunication
SendGrid
Transactional email
Email
SMTP
Standard email protocol
Email
Telegram
Bot messaging
Messaging
Slack
Team messaging
MessagingStorage & Database
Files
S3, Azure Blob, GCS, local
File storage
PostgreSQL
Relational database
Database
Redis
In-memory cache
CacheCloud Providers
AWS
Amazon Web Services
Cloud services
Azure
Microsoft Azure
Cloud services
GCP
Google Cloud Platform
Cloud servicesDevOps
Docker
Container runtime
Containers
GitHub
Git & CI/CD
Code and CIBusiness Tools
Stripe
Payments
Payments
Zapier
Workflow automation
Automation
HubSpot
CRM
CRM
Notion
Documentation
DocumentationFLEXIBLE ARCHITECTURE
What happens if Huitzo doesn't have a built-in integration for my tool?
Ship your packs with exactly the integrations you need, or none at all.
Ship Without
Pure Logic Packs
Build Intelligence Packs that contain pure business logic with no external dependencies. Perfect for data transformations, calculations, and algorithmic workflows.
Ship With Built-in
Declare and Use
Declare required integrations in your pack manifest. Configure credentials once in the platform, and the runtime passes them to the driver at execution. Your code uses the context object.
Add Custom Connectors
Extend the Runtime
Need to connect to a proprietary system? Licensed deployments can add a custom connector, built by the partner or your engineering team against the Huitzo SDK and docs.
CUSTOM CONNECTORS
Can I build a custom connector for a proprietary internal system?
Huitzo's platform separates your AI logic from integration details. A custom connector is development work: the partner or your own engineering team builds it against the Huitzo SDK and docs, then registers it for any Intelligence Pack to call.
- Connect to legacy internal systems (SAP, Oracle, mainframes)
- Integrate proprietary algorithms (Risk models, pricing engines)
- Support future hardware (Specialized accelerators, IoT devices)
- Deploy on your infrastructure (On-premises, private cloud)
QUESTIONS
Integration questions
Does Huitzo work with OpenAI?
Can I add a custom integration?
What protocols are supported out of the box?
Does Huitzo run inside an air-gapped environment?
Can I integrate with my existing identity provider?
How do Intelligence Packs share state across integrations?
AI operating system for regulated industries
Simple by design. Infinitely scalable. Runs where your data lives.
- Product access
- huitzo.ai, the hosted marketplace, is open by invitation to teams evaluating the product.
Or email [email protected] directly.