Topics Covered:

Provider Architecture: OpenAPI-based external service integrations
Connector Management: Authentication and access control for providers
LLM Models: Configure and manage available LLM models
Central Repository: Curated provider library with sync capabilities
OAuth and authentication flows
Provider versioning and updates

Provider Architecture

Providers in EverWorker represent external APIs or service ecosystems (e.g., Microsoft 365, OpenAI, GitHub) defined using OpenAPI specifications. Each Provider includes a detailed schema of endpoints and methods, enabling Workers to interact with external services programmatically and consistently.

Providers serve as the foundation for all tool-based capabilities within Workers
Managed centrally by EverWorker through an automated API Scraper system that keeps specs current
Can support hundreds of providers, covering major enterprise and SaaS ecosystems
Supports both JSON and YAML formats for OpenAPI specification uploads

Connector Management

Connectors are instances of Providers, configured with authentication credentials and context-specific settings (e.g., for different tenants, users, or environments).

Multiple Connectors can be created per Provider (e.g., two Gmail accounts, separate OpenAI keys)
Support for authentication parameters such as API keys, OAuth tokens, tenant IDs, and more
Assigned on a per-node or per-worker basis in AI Workflows, and per skill in AI Workers
Configurable via a secure UI, with role-based access (Builders/Admins)

LLM Models

LLM Models allows Admins to configure and manage the Large Language Models available across the platform. This feature is accessible as a tab within Providers in the navigation menu (alongside Your Connectors, Your API Providers, MCP Servers, and Incoming Webhooks).

Key Capabilities:

Add New LLM Models: Register new models with their provider, capabilities, and pricing
Configure Model Features: Define what each model supports:
- Reasoning, Streaming, Structured Output
- Function Calling, Fine-tuning, Realtime
- Predicted Outputs
Input/Output Configuration:
- Maximum token limits for input and output
- Cost per million tokens (CPM) for billing/budgeting
- Cached input pricing
- Modality support (text, image, audio)
Enable/Disable Models: Toggle model availability across the platform
Edit and Delete: Update model configurations or remove deprecated models

Model Properties:

Property	Description
Name	The model identifier (e.g., gpt-4, claude-3-sonnet)
Provider	The LLM provider (e.g., openai, anthropic, google)
Status	Active or Inactive
Features	Capabilities like streaming, function calling, etc.
Input Tokens	Maximum context window size
Output Tokens	Maximum response length
Cost (CPM)	Pricing per million tokens for input and output

Central Repository

EverWorker maintains a curated library of Providers, regularly updated through the API Scraper engine. This is known as the Provider Central Repository.

Customers select and sync only the Providers relevant to their needs
Sync operation pulls the latest, validated OpenAPI definitions
Eliminates manual setup and version drift
Available during initial setup or reconfigured on demand

OAuth and Authentication Flows

EverWorker supports multiple authentication models to handle secure, multi-user environments:

User OAuth: Requires end-user login and consent (used when Workers need access to user's personal accounts)
App Token: Uses system-level access credential.
Supports OAuth scopes, redirect flows, and token refresh logic
Integrated seamlessly in chat, connector config UI, and session workflows

Provider Versioning and Updates

Providers are versioned through EverWorker's backend tooling:

Admins and Builders are notified of available updates and can sync new versions from the Central Repository
Version history and audit logs are maintained for governance and debugging

Summary

The Provider system in EverWorker ensures secure, flexible, and up-to-date integrations across any enterprise or third-party services. With a clear separation between Providers and Connectors, customers can manage complex authentication needs while safely reusing integrations across Workers. Combined with a central repository and real-time version tracking, EverWorker makes external service orchestration fast, reliable, and enterprise-ready.