Hackeet

A Swiss Army knife for Data and Agent Engineers.

Visually build high-frequency data pipelines, automated workflows, or create production-ready AI agents, all running on your own hardware.

Animated Hackeet flow diagram
Book a demo See the Hedy case study
  • No SaaS
  • No credit card
  • Easy install
  • Private by design

One platform, many uses

[01] The technology

The power of code-based agents, with the flexibility of visual programming.

Hackeet builds production-ready systems with the performance of code-based agents.

  • 01

    Visible logic

    Agent and automation logic that everyone can see and understand.

  • 02

    LangChain performance

    The same performance as LangChain, the agent framework running underneath.

  • 03

    Built for performance

    Process large volumes of data fast, or run multi-tool agents through complex tasks.

Example Hackeet workflow: connect, Pinecone store and search, vector store, OpenAI embeddings and error logging
[02] Develop & deploy

Develop and deploy your way

Hackeet fits into your existing tools and runs wherever you need it.

  • 01

    Native Git

    Native Git integration, no database: Hackeet files are source files like any other.

  • 02

    Extensible components

    Extensible tools and components that connect to your systems, built in Hackeet or as DLLs with the C/C++ development kit.

  • 03

    Your infrastructure

    Deploy as a standard web server, an MCP server, or embedded in your Python, C/C++, JavaScript or C programs. Rust and Go coming soon.

  • 04

    Claude Code

    Claude Code integration to develop in Hackeet.

// your_libraries/options.cpp
double black_scholes_price(OptionType type, double spot, double strike,
                            double rate, double dividend_yield,
                            double volatility, double maturity) {
    ...
    return price;
}

↓ Becomes a component

OptionPricer
("put", 100.0, 105.0, 0.02, 0.0, 0.25, 365) → price
Web servers

Web servers

Expose your flows as a standard web API.

MCP servers

MCP servers

Make your tools available to any MCP-compatible AI client.

RAG

RAG

Connect your documents and databases to your AI agents.

Embedded Applications

Embedded Applications

Embed Hackeet and turn your legacy C++, JavaScript, and Python code into high-performance modules and intelligent AI-powered applications.

[03] Components

Ready-to-use components

Every connector is a visual component, ready to drop into your flows.

...And many more

[04] Under the hood

Private by design, light by nature

Software, not SaaS. Hackeet installs on your side and adapts to your architecture, from a central server to an embedded device.

Offline

Private by design

  • Runs without an internet connection
  • No data transfer
  • You own your code, GDPR compliant
  • No automatic updates
Light

Lightweight

  • Low memory footprint
  • Light deployment possible: one server per workstation, no central server
Low-level

Low-level & embedded

  • Low-level hardware connection
  • Embedded deployment, all the way to robotics
[05] Case study

Hedy, a legal AI agent

Hedy is a RAG-powered AI agent built for the legal sector.

A few weeks part-time, from concept to MVP
hedy.hkt run log · 4 steps
RAG · the firm's internal resources OK
MCP/API · French and European law OK
OCR · reading attached documents OK
LLM · chatbot, model of your choice RUNNING
[06] Pricing

One license, options to match your needs

License pricing depends on the size of your company. Add only the options you need.

The core

Hackeet annual license

Custom quote, based on your company size

  • The platform and every available component
  • Deploy on your workstations, your servers or embedded
  • Your data stays with you, no connection required
  • Updates installed when you decide
Get a quote
+ Option

Maintenance

Technical support and platform maintenance over time.

+ Option

Expert engagement

Our team works with yours to design and deploy your projects.