Content
# Roampal
[](https://roampal.ai)
[](https://www.python.org/downloads/)
[](https://tauri.app/)
[](https://roampal.ai)
[](LICENSE)
**Memory that learns what works.**
*Say it worked. Say it didn't. The AI remembers.*
Stop re-explaining yourself every conversation. Roampal remembers outcomes, learns from feedback, and gets smarter over time—all 100% private and local.
<p align="center">
<img src="screenshots/whatwebeenupto.png" alt="Roampal - AI Chat with Persistent Memory" width="800">
</p>
> **85.8% non-adversarial on LoCoMo (1,986 questions). +23 pts over raw ingestion. Absorbs 1,135 poison memories losing only 4 pts.** [(Paper)](https://github.com/roampal-ai/roampal-labs)
<p align="center">
<a href="https://github.com/roampal-ai/roampal">
<img src="https://img.shields.io/github/stars/roampal-ai/roampal?style=social" alt="GitHub Stars">
</a>
</p>
---
## Benchmark Results
*LoCoMo dataset (1,986 questions, 5 categories, corrected ground truths). Evaluated with [roampal-labs](https://github.com/roampal-ai/roampal-labs). Dual-graded by local 20B + MiniMax M2.7.*
| Metric | Result |
|--------|--------|
| **Non-adversarial accuracy (MiniMax-regraded)** | **85.8%** |
| **Overall (all 5 categories)** | **76.6%** |
| **vs raw ingestion baseline** | **+23 pts** (76.6% vs 53.0%, p<0.0001) |
| **Poison resilience** | **-4.2 pts** after 1,135 adversarial memories |
| **No-memory baseline** | 6.0% (model has zero LoCoMo knowledge) |
| **Architecture vs model** | Architecture: +23 pts. Model swap (GPT-4o-mini): 1.5-2.5 pts |
- System learns through natural conversation, not transcript ingestion
- Absorbs 1,135 poison memories with spoofed trust signals, retaining 72.4% accuracy
- Wilson scoring hurts retrieval at every stage (p<0.001) — removed from ranking
<details>
<summary>Component-level retrieval ablation</summary>
| Config | Hit@1 Clean | Hit@1 Poison | p-value |
|--------|-------------|--------------|---------|
| **TagCascade + cosine** | **27.3%** | **29.0%** | **baseline** |
| Overlap + cosine | 25.8% | 28.0% | p=0.0003 |
| Pure CE | 25.4% | 28.4% | — |
| TagCascade + Wilson | 23.0% | 25.0% | p<0.0001 |
- Cross-encoder: +17.8 Hit@1 over cosine (p<0.0001)
- Tag routing (two-lane): +6.1 Hit@1 clean, +7.5 poison (p<0.0001)
- Wilson: -4.3 Hit@1 in every configuration
- Nursery slot: zero benefit (p=1.0)
Full methodology in [roampal-labs](https://github.com/roampal-ai/roampal-labs)
</details>
---
## Quick Start
1. **[Download from roampal.ai](https://roampal.ai)** and extract
2. Install [Ollama](https://ollama.com) or [LM Studio](https://lmstudio.ai)
3. Right-click `Roampal.exe` → **Run as administrator**
4. Download a model in the UI → Start chatting!
Your AI starts learning about you immediately.
---
## Table of Contents
- [Benchmark Results](#benchmark-results)
- [Key Features](#key-features)
- [MCP Integration](#mcp-integration)
- [Architecture](#architecture)
- [Supported Models](#supported-models)
- [Documentation](#documentation)
- [Pricing](#pricing)
---
## Key Features
**Memory That Learns**
- Outcome tracking: Scores every result (+0.2 worked, -0.3 failed)
- Smart promotion: Good advice becomes permanent, bad advice auto-deletes
- Cross-conversation: Recalls from ALL past chats
**Your Knowledge Base**
- Memory Bank: Permanent storage of preferences, identity, goals
- Books: Upload .txt/.md docs as searchable reference
- Pattern recognition: Detects what works across conversations
**Privacy First**
- 100% local: All data on your machine
- Works offline: No internet after model download
- No telemetry: Your data never leaves your computer
---
## MCP Integration
Connect Roampal to **Claude Desktop, Cursor**, and other MCP-compatible tools.
```
Settings → Integrations → Connect → Restart your tool
```
**7 tools available**: `search_memory`, `add_to_memory_bank`, `update_memory`, `archive_memory`, `get_context_insights`, `record_response`, `score_memories`
[Full MCP documentation →](dev/docs/architecture.md#mcp-integration)
---
## Architecture
```
┌─────────────────────────────────────────────────────────┐
│ 5-TIER MEMORY │
├─────────────┬─────────────┬─────────────┬──────────────┤
│ Books │ Working │ History │ Patterns │
│ (permanent) │ (24h) │ (30 days) │ (permanent) │
├─────────────┴─────────────┴─────────────┴──────────────┤
│ Memory Bank │
│ (permanent user identity/prefs) │
└─────────────────────────────────────────────────────────┘
```
**Core Technology:**
- TagCascade Retrieval: Tag-routed search + cross-encoder reranking (ONNX)
- Outcome-Based Learning: Memories adapt based on feedback
- Sidecar LLM: Background model summarizes exchanges, extracts facts and tags
[Architecture deep-dive →](dev/docs/architecture.md)
---
## Supported Models
Works with any tool-calling model via Ollama or LM Studio:
| Model | Provider | Parameters |
|-------|----------|------------|
| Llama 3.x | Meta | 3B - 70B |
| Qwen 2.5 | Alibaba | 3B - 72B |
| Mistral/Mixtral | Mistral AI | 7B - 8x22B |
| GPT-OSS | OpenAI (Apache 2.0) | 20B - 120B |
---
## Documentation
| Document | Description |
|----------|-------------|
| [Architecture](dev/docs/architecture.md) | 5-tier memory, knowledge graphs, technical deep-dive |
| [Benchmarks](dev/docs/releases/v0.3.1/RELEASE_NOTES.md#benchmark-evidence) | LoCoMo evaluation, TagCascade results |
| [Release Notes](dev/docs/releases/v0.3.1/RELEASE_NOTES.md) | Latest: TagCascade Retrieval, Sidecar LLM, ONNX CE, Two-Lane Injection |
---
## Important Notices
**AI Safety**: LLMs may generate incorrect information. Always verify critical information. Don't rely on AI for medical, legal, or financial advice.
**Model Licenses**: Downloaded models (Llama, Qwen, etc.) have their own licenses. Review before commercial use.
---
## Support
- **Discord**: https://discord.gg/F87za86R3v
- **Email**: roampal@protonmail.com
- **GitHub**: https://github.com/roampal-ai/roampal/issues
- **Author**: [Logan Teague](https://www.linkedin.com/in/logan-teague-6909901a5/)
---
## Pricing
**Free & open-source** (Apache 2.0 License)
- Build from source → completely free
- Pre-built executable: **$9.99 one-time** (saves hours of setup)
- Zero telemetry, full data ownership
---
**Made with love for people who want AI that actually remembers**
