A shared data platform your AI agents can query concurrently.
DuckDB is great in-process, but agents working together need a shared, concurrent data platform. Hardcharger is a lightweight, high-performance data platform that speaks Warp SQL, so many LLM agents query at once — with a Python ADBC driver that drops into LangChain, CrewAI, and your own frameworks.
Apache-2.0. Every feature included. No license keys or per-vCPU software fees.
$ HARDCHARGER_ADMIN_PASSWORD=change-me hardcharger-server --database agents.duckdbpy> import adbc_driver_hardcharger.dbapi as hcpy> db = hc.connect(uri="hardcharger://localhost:31337?tls=false", username="hardcharger", password="change-me")agent_a ── query ─┐agent_b ── query ─┼─▶ one shared platformagent_c ── query ─┘
Concurrent access
Multiple agents query the same data platform at the same time over Warp SQL.
Drops into your framework
Python ADBC driver for LangChain, CrewAI, and whatever you are building. Plus an MCP server for Claude Desktop and Claude Code.
Secure by default
TLS, SSO, roles, row policies, column masking and tenants are included under Apache-2.0.
Install Hardcharger, free
Native binaries or Docker. Start the data platform in minutes.
Point it at your data
Parquet, Iceberg, Delta, or DuckLake on your object storage. No copies.
Connect your agents
One ADBC connection string, shared across every agent in the workflow.
Hardcharger for AI agents — FAQ
Why not just use DuckDB in-process for my agents?
In-process DuckDB binds the database to a single process, so concurrent agents fight over file locks or need fragile copies. Hardcharger wraps DuckDB in a server that speaks Warp SQL, giving every agent shared, concurrent access to one platform — with TLS and authentication built in.
How do my agents connect to Hardcharger?
Hardcharger uses Warp SQL over HTTP, with native ADBC, JDBC and JavaScript clients. Use dbt-hardcharger for dbt and the Hardcharger MCP server for AI assistants. Check SQL dialect differences and client compatibility when migrating.
Is Hardcharger secure enough for production agent workloads?
Hardcharger is Apache-2.0 with every feature included: access control, SSO, audit logging and the web UI. There are no editions, license keys or per-vCPU software fees. You pay for your infrastructure and any optional managed services or support.
What does Hardcharger cost?
Our 1 TB workload derived from TPC-H ran all 22 queries in 42.4 seconds for $0.09 of AWS compute on one machine with Hardcharger 2.0.0. Infrastructure, storage and managed-service costs vary; measure your own workloads. Hardcharger is Apache-2.0 with every feature included: access control, SSO, audit logging and the web UI. There are no editions, license keys or per-vCPU software fees. You pay for your infrastructure and any optional managed services or support.
Give every agent in the workflow one fast, shared data platform.
Apache-2.0. Every feature included. No license keys or per-vCPU software fees.