Documentation

State schemas, engines, persistence, and graph-building foundations for the Haive runtime.

Production-ready agent implementations, reasoning patterns, memory, and 20+ RAG methodologies.

Search, calculators, APIs, and integrations for composable agent tooling.

LLM-powered game agents and benchmarking environments across 20+ competitive and social settings.

Starter configurations, recipes, and domain-specific deployments for common enterprise use cases.

ACP and MCP deployment framework with serializable config, WebSockets, and Postgres-backed persistence.

Dynamic MCP integration enabling agents to discover, install, and use external tools with approval gates.

MCP-style protocol for orchestrating single-agent services and multi-agent graphs over JSON-RPC.

Database intelligence agent for DB-RAG, schema-aware exploration, guarded SQL execution, and iterative self-improvement.

Unified quantitative and market-systems platform for forecasting, stochastic simulation, risk modeling, derivatives, backtesting, and agent-facing financial workflows.

Multi-agent skills registry for curating, normalizing, syncing, and distributing capabilities across Codex, Claude Code, Cursor, and agent runtimes.

Pydantic-based reasoning toolkit for symbolic logic, proof validation, SAT workflows, and machine-checkable traces.

Guide of 200+ adversarial prompt cases with mapped defenses, detection rules, and workflow integrations for agent guardrails.

Open-source Perplexity-style research engine with quick search, pro search, and deep research workflows for planning, retrieval, reflection, and synthesis.

Agentic OpenAPI-to-MCP system that turns API specifications into FastMCP and LangGraph server artifacts with improved tool documentation and prompts.

Database-backed prompt registry and runtime delivery layer for versioning, storing, and serving templates, chains, and few-shot sets across agent frameworks.

Unified LLM integration layer for LangChain-first applications with provider-aware model parsing, typed settings, live model discovery, and cost helpers.

Converts handwritten notes, PDFs, and images into structured LaTeX with OpenCV preprocessing, OCR, and LLM-based self-correction.

Runtime structural typing and shape algebra across Pydantic, dataclasses, TypedDict, attrs, and protocol-style interfaces.

Unified Python debugging, profiling, tracing, and structured logging toolkit for richer runtime visibility and faster issue diagnosis.

CLI and Python library for cleaning repositories of secrets and sensitive files before sharing or open-sourcing them.

Typed Python wrapper for Private Internet Access piactl with environment settings, smart region selection, and proxy adapters.

Python fetching and retrieval toolkit with documented interfaces for extraction, transport, and utility workflows.

FastMCP integration around Tavily search workflows for agent toolchains and structured retrieval surfaces.

MCP server and tooling layer for Polymarket data access and agent-facing prediction-market workflows.

Older but still interesting Django automation toolkit with live docs, kept in the archive tier rather than the flagship set.