For Developers
Integrate the only AI biblical research tool backed by a dedicated vector knowledge base.
Unlike generic AI that relies on training data and can hallucinate, BEREAN.AI maintains a dedicated vector database of 1.6 million indexed scholarly passages. Every answer is grounded in retrieval-augmented generation (RAG) that searches real sources first. The AI never guesses — it searches, retrieves, and cites.
The Vector Knowledge Base
Tier 1 Factual & Linguistic Data
Objective, scholarly data presented as established fact. Includes:
Tier 2 Interpretive Sources (always attributed)
Scholarly commentary and theology, always presented with tradition attribution. Includes:
REST API
Two modes, both returning structured JSON. Free to use with fair-use limits.
Rate Limit: 20 requests per day per IP address. The limit resets every 24 hours. When you hit the limit, the API returns a clear JSON error message explaining the limit — your integration is not broken. Need higher limits for a production app? Contact [email protected].
Quick pastoral/theological answer (250-400 words). Best for life questions, doctrine overviews, practical faith guidance.
curl -X POST https://berean.ai/api/v1/question \
-H "Content-Type: application/json" \
-d '{"question": "What does the Bible say about anxiety?"}'
Deep academic answer with RAG search across the full vector database. Returns the answer plus source metadata. Best for Greek/Hebrew word studies, cross-tradition theology, exegetical questions.
curl -X POST https://berean.ai/api/v1/scholar \
-H "Content-Type: application/json" \
-d '{"question": "What does the Greek word agape mean in John?"}'
Response includes:
{
"mode": "scholar",
"question": "What does the Greek word agape mean in John?",
"answer": "The Greek noun agapē (ἀγάπη) in John...",
"sources_count": 30,
"sources": [
{ "tier": 1, "collection": "lexicons", "source": "Thayer's Greek Lexicon", "text": "..." },
{ "tier": 2, "collection": "commentaries", "source": "Calvin", "tradition": "Reformed", "text": "..." }
],
"model": "gemini-3.1-flash-lite",
"elapsed_ms": 8500
}
OpenAPI 3.1 spec for ChatGPT Custom GPT Actions or any OpenAPI-compatible client.
Basic API information and available modes.
ChatGPT Integration
BEREAN.AI is available as a Custom GPT. Users can access it directly in ChatGPT on any device:
To build your own Custom GPT using BEREAN.AI, import the OpenAPI spec from https://berean.ai/api/v1/openapi.json in the GPT Builder Actions tab.
MCP Server
A live Model Context Protocol server for Claude Desktop, Cursor, and other MCP-aware AI clients. Streamable HTTP transport at https://berean.ai/mcp.
Available Tools
ask_question — Pastoral/theological answer (250-400 words) grounded in Reformed theology with Scripture references.
scholar_query — Academic research powered by RAG search across 1.6M+ indexed passages. Returns detailed answers with source citations from lexicons, commentaries, church fathers, and more.
search_sources — Search the vector database directly. Returns raw source passages from the knowledge base without generating an AI answer. Useful for getting primary source data.
Connect from Claude Desktop
Add this to your claude_desktop_config.json:
{
"mcpServers": {
"berean": {
"url": "https://berean.ai/mcp"
}
}
}
Connect from Cursor
Add to your project's .cursor/mcp.json:
{
"mcpServers": {
"berean": {
"url": "https://berean.ai/mcp"
}
}
}
Connect from any MCP client
Point your Streamable HTTP MCP client to:
https://berean.ai/mcp
The server supports session management, SSE streaming, and all standard MCP protocol features. Rate limited to 20 tool calls per day per IP (same as the REST API). Initialize/session requests are not counted against the limit.
Why This Matters
Every AI hallucinates about the Bible. BEREAN.AI doesn't — it searches 1.6 million indexed passages from real lexicons, commentaries, and manuscripts before answering. When it cites Calvin, it's quoting from an actual embedded copy of Calvin's 45-volume commentary set. When it analyzes Greek morphology, it's pulling from MorphGNT and MACULA word-level syntax data. When it presents the Orthodox perspective, it's drawing from Chrysostom's actual homilies in the NPNF collection.
This isn't a wrapper around ChatGPT. It's a dedicated retrieval-augmented generation pipeline with a curated, multi-tradition scholarly knowledge base that no other biblical AI tool has.