FastAIMemory 0.1.0 [ALPHA-2026-06-14] — Unified Conversation History and Memory Orchestration for Java

---

💡 Extremely lightweight, provider-independent, thread-safe conversation history, formatters, and memory-trimming utilities for the FastJava AI Ecosystem.

FastAIMemory is a primitive context manager for Java. It stores conversation messages, optimizes history boundaries (sliding windows, character limits, token estimates), and converts structural data into target-formatted prompt strings (ChatML, Gemini, Claude) to keep LLM context clean and safe.

---

Quick Start

import fastaimemory.*;

public class Demo {
    public static void main(String[] args) {
        // 1. Initialize thread-safe conversation history
        ConversationHistory history = new ConversationHistory();
        history.system("You are a helpful coding assistant.");
        
        history.user("Hello!");
        history.assistant("Hi! How can I help you today?");
        
        history.user("Write a quicksort in Java:");

        // 2. Format history dynamically for different providers
        MemoryContextBuilder builder = new MemoryContextBuilder(new ChatMLFormatter());
        String chatMLPrompt = builder.build(history);
        System.out.println(chatMLPrompt);
    }
}

---

Why FastAIMemory?

• Provider Agnostic — Decoupled from specific provider APIs. Works seamlessly with OpenAI, Gemini, Claude, Ollama, and local runtimes.
• Thread Safe — Thread-safe ConversationHistory using lock synchronization makes it reliable for concurrent multi-agent environments.
• Polymorphic Formatters — Implement MemoryFormatter to structure output prompts dynamically using plain text, ChatML, or specialized provider tokens.
• Zero Dependencies — Built on pure Java 17, keeping compiling speeds lighting fast and jar footprint zero-bloat.

---

Installation

Add the JitPack repository and the dependency to your pom.xml:

<repositories>
    <repository>
        <id>jitpack.io</id>
        <url>https://jitpack.io</url>
    </repository>
</repositories>

<dependencies>
    <dependency>
        <groupId>com.github.andrestubbe</groupId>
        <artifactId>FastAIMemory</artifactId>
        <version>0.1.0</version>
    </dependency>
</dependencies>

---

API Reference

ConversationHistory

Thread-safe container storing conversation turns.

• add(ConversationRole role, String text): Appends a raw message.
• system(String text) / user(String text) / assistant(String text): Helper methods.
• messages(): Returns a read-only list copy of the messages.
• clear(): Wipes history.

MemoryWindow

Utility for trimming historical context prior to sending requests:

• trimToMessages(List<ConversationMessage> messages, int maxMessages): Retains the first SYSTEM message, keeping only the latest $N$ messages.
• trimToCharacters(List<ConversationMessage> messages, int maxChars): Trims oldest messages until text length fits within limits.
• trimToEstimatedTokens(List<ConversationMessage> messages, int maxTokens): Uses a $chars / 4$ heuristic estimation for context pruning.

---

Related Projects

• FastAI - Unified AI client interface for Java
• FastAIModel - Native local inference runtime (GGUF/ONNX)
• FastCore - Unified JNI loader and platform abstraction

---

License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Part of the FastJava Ecosystem — Making the JVM faster. Small package. Maximum speed. Zero bloat. 🚀📋