Leviathan

Leviathan is a search tool that indexes large datasets into ranked full-text records for agents.

Leviathan is a single static binary that indexes large datasets like JSON, CSV, and SQLite into a ranked full-text search index. Agents ask questions in plain words and receive short, cited result cards instead of reading raw history. It uses FTS5 and BM25 ranking to return relevant records with low latency and minimal token usage. The tool supports scoped searches, time filters, and dynamic updates via upsert and delete commands. It integrates with any agent via a CLI skill or through an optional Model Context Protocol server.

Leviathan at a glance

Best for: Software engineer, Researcher
Runs on: CLI, self-hosted, open source
Pricing: Free
Released: October 2026
Category: AI research tool
Website: github.com/elstongun/leviathan

Key features

Index JSON, JSONL, CSV, TSV, and SQLite data
Return short, cited result cards for agent queries
Run full-text search with BM25 ranking and boosts
Filter results by date, fields, and groups
Update indices atomically with upsert and delete
Serve read-only tools via Model Context Protocol
Operate as a single static binary with no runtime dependencies

Use it for

Query large maintenance logs efficiently with AI agents
Retrieve specific records from database exports using plain text
Search historical datasets while minimizing token consumption

Alternatives to Leviathan

Exa — Provides high-accuracy web search and deep research endpoints for AI agents.
Firecrawl — Provides direct web data access for AI agents with open source flexibility.
SearchCLI — Provides a stable, reviewable CLI for managing Volcengine AI search workflows.
HyperResearch — Generates verified reports with citations checked against full source texts and a persistent vault.
past.dev — Provides agents with accurate, timestamped context instead of static retrieval.
LLMSchema — Provides AI coding agents with accurate database schemas without context bloat.
RAGFlow — Provides an integrated platform combining data processing with agent workflows.
seekdb — Unifies relational, vector, and text data for production-grade AI agents.
LLM Wiki Research — Compiles parallel agent research into source-backed markdown knowledge bases.
Searchbase — Independent web access with support for REST, MCP, self-hosting, and managed cloud.
Limma — It answers queries with complete sets and proof instead of partial search fragments.
Lune — Accesses trusted knowledge from top conference papers for AI agents.

All Leviathan alternatives