Frameworks & Tools

Open-source building blocks for reliable, transparent AI agents.

HANERMA

v1.0.1 • pip install hanerma

Hierarchical Atomic Nested External Reasoning & Memory Architecture.
Open-source framework focused on perfect long-term memory, atomic-verified reasoning, intelligent routing, and calibrated confidence for AI agents.

Verified Features

• Auto Mode (zero-shot classifier selects direct/reasoning/memory)
• Force Memory / Force Reasoning / Force Both overrides
• Direct Mode (bypasses overhead for speed)
• Per-query mode override
• Persistent memory recall across turns
• JSON save/load for memory store
• Full internal logging (ai.logs)

Core Architecture

• Atomic Reasoning: Decomposes to verifiable units + bottom-up synthesis
• External Memory: sentence-transformers embeddings (unlimited history)
• Intelligent Routing: Zero-shot path selection
• Calibrated Confidence: Every answer includes Confidence: X%

Installation

pip install hanerma

Requires Python ≥ 3.8

Basic Usage

from hanerma import HANERMA

ai = HANERMA(api_key="sk-or-your-key", model="meta-llama/llama-3.1-70b")
print(ai.ask("What is 17×23?"))
print(ai.ask("What was my previous question?"))

Install via PyPI Full Documentation GitHub