MechaMental Documentation
Build, deploy, and manage production-grade AI applications with MechaMental.
MechaMental is a full-stack AI orchestration platform for building deterministic, versioned, production-grade AI workflows. This documentation covers everything from your first app to advanced platform administration.
Quick Start
Core Concepts
Architecture
Organization, workspace, and app hierarchy.
Pipelines
Multi-stage processing chains with 13 step types.
Endpoints
Expose pipelines as HTTP APIs with auth and streaming.
Environments
Production, Staging, and Development configurations.
Releases
Immutable deployment snapshots with rollback.
Models
Logical model abstraction with multi-provider routing.
Knowledge
Namespaces, sources, threads, memory, artifacts, and vault.
Tools
MCP-compatible integrations and external services.
Cortex
Built-in chat interface with bundles and scratchpad.
Security
RBAC, SSO/OIDC, vault secrets, audit logs, prompt protection.
Guides
Building a Pipeline
End-to-end walkthrough of creating an augmentation pipeline.
Configuring Step Types
Deep dive into all 13 step types and their configuration.
Managing Sources
Upload files, configure chunking, and monitor processing.
Working with Tools
Browse the library, install instances, and use in pipelines.
Deploying a Release
Create releases, deploy to environments, and rollback.
API Tokens
Create, use, and manage API tokens for programmatic access.
Administration
IAM & Access Control
Roles, policies, and SSO/OIDC configuration.
Model Configuration
Provider accounts, logical models, and entitlements.
Audit Logs
Immutable event log with filtering and export.
Billing & Limits
Usage dashboard, spending limits, and optimization.
API Reference
Browse the auto-generated API Reference with request/response schemas for all 389 endpoints across 47 tags.