Introduction

Inherent is managed RAG infrastructure. It gives your AI application a single, versioned knowledge base backed by PostgreSQL for truth and Weaviate for semantic search -- with every retrieval logged for audit.

You send documents in. You get grounded context out. No vector-DB tuning, no chunking pipeline to maintain, no "it worked yesterday" debugging sessions.

The Problem

Building a RAG pipeline that works is easy. Building one that works reliably in production is not:

• Stale context -- your AI serves outdated information because nothing tracks document freshness.
• Irreproducible results -- same query, different answers. Good luck debugging that without retrieval logs.
• No separation of concerns -- your authoritative data and your search index are tangled together. When one breaks, both break.
• Multi-tenancy is an afterthought -- isolating knowledge per customer means re-inventing workspace scoping on every project.

Inherent solves all four.

How It Works

Inherent exposes a single REST API that handles ingestion, retrieval, and audit. Under the hood, it separates truth from memory:

┌──────────────────────────────────────────────────────────┐
│                   Your AI Application                    │
└──────────────────────────────────────────────────────────┘
                            │
                    Inherent Public API
                            │
          ┌─────────────────┼──────────────────┐
          ▼                 ▼                  ▼
   ┌────────────┐    ┌────────────┐    ┌────────────────┐
   │   Truth    │    │   Memory   │    │   Audit Trail  │
   │ PostgreSQL │    │  Weaviate  │    │   (per query)  │
   └────────────┘    └────────────┘    └────────────────┘

Truth Layer (PostgreSQL)

The authoritative store for every document, version, and chunk. When you upload a file, Inherent parses it, splits it into chunks, and writes the structured data here. This is the single source of truth -- if the search index is wiped, it can be rebuilt from this layer.

Memory Layer (Weaviate)

Stores vector embeddings of every chunk for fast semantic search. When your app queries Inherent, the Memory Layer finds the most relevant chunks, and the Truth Layer hydrates them with full metadata. Two systems, one API call.

Audit Trail

Every search request is logged: what query was sent, which chunks were returned, the scores, and the timestamp. When a user reports a wrong answer from your AI, you can trace exactly what context it received and when.

Key Features

Managed Ingestion Pipeline

Upload a PDF, Markdown file, DOCX, or plain text via the API. Inherent handles parsing, chunking, embedding, and indexing asynchronously. No Airflow DAGs. No Lambda functions. No glue code.

Deterministic Retrieval

Same query, same documents, same results. Inherent's retrieval is reproducible by design, which means your tests and debugging sessions actually work.

Multi-Tenant Workspaces

Each workspace is fully isolated -- separate documents, separate search index, separate API keys. Built for B2B SaaS teams that need per-customer knowledge contexts without building tenant isolation from scratch.

Full Audit Trail

Every retrieval is recorded with query text, returned chunks, relevance scores, and timestamps. Required for regulated industries. Useful for everyone else.

Use Cases

• RAG chatbots -- ground your chatbot in up-to-date company docs with reliable context retrieval.
• AI copilots -- feed your copilot the right internal knowledge without building a retrieval pipeline.
• Document Q&A -- let users ask questions against uploaded files with cited, traceable answers.
• Knowledge management for regulated industries -- healthcare, legal, and finance teams that need audit trails on every piece of context their AI consumes.

Quick Links

Quickstart	Upload a document and search it in 5 minutes
Authentication Guide	API keys and security
Uploading Documents	Supported file types, async processing
Searching	Search parameters and best practices
Retrieving Context	Chunks, context assembly for LLMs
API Reference	Complete endpoint documentation

Pricing

Plan	Price	Highlights
Starter	$149/mo	500 documents, 10K API calls/mo, 3 workspaces, audit trail
Pro	$349/mo	Unlimited documents, 100K API calls/mo, 10 workspaces, chunk editing, MCP server
Business	$799/mo	Everything in Pro, 500K API calls/mo, unlimited workspaces, SSO, SLA
Enterprise	$2,000+/mo	Custom limits, dedicated infrastructure, priority support, custom SLAs

All plans include the managed ingestion pipeline, semantic search, and the full audit trail.

Get started ->