open-dream-rsi

Open Dream-RSI

What is this?

Every AI agent pays the same tax over and over: it fails a task, works out

what went wrong, and then throws that understanding away when the session

ends. Open Dream-RSI keeps what was learned — without touching the model.

The LLM stays frozen. What improves is everything around it:

• the exploration policy (even re-written as Python code by the LLM

itself, gated by replay before it is trusted),

• a library of verified solutions (recipes) reused as warm starts,
• and the world model the policy is scored against: a Discovery Tree of

every attempt, enriched with replay-safe error-class facts by the Sentinel

engine.

Between runs the system "dreams": it replays thousands of strategies over

the recorded history in-process, at zero API cost, and keeps only what

beats the incumbent. An open implementation of *Recursive Self-Improvement

through Evolving Worlds* (Dream-RSI, Zheng et al.,

[arXiv:2609.14858](https://arxiv.org/abs/2609.14858)).

Zero runtime dependencies — Python 3.10+ stdlib only, ~8.6k lines of

library code, ~4.4k lines of tests, no build step.

Not another memory plugin. Experience-memory systems record what the

agent learned and reuse it as-is. ODR is an evolving worlds loop: every

attempt is recorded into a Discovery Tree, and between runs the system

dreams — replaying thousands of strategy variants against that recorded

world. Nothing is promoted on the model's word alone: a candidate policy or

recipe replaces the incumbent only after surviving counterfactual replay.

How it compares — stored state, how entries earn trust, runtime cost,

against Mem0, Letta, Reflexion-style reflection, DSPy optimizers and skill

libraries: [docs/comparison.md](docs/comparison.md).

> 📄 Preprint: *Open Dream-RSI: An Open-Source Library for Recursive

> Self-Improvement Around a Frozen LLM, with Replay-Gated Learned Policies and

> a Curated Knowledge Base* — P. Orwat, 2026.

> [PDF](paper/main.pdf) · [LaTeX source](paper/main.tex)

---

Quick start

| Agent | Install |

|---|---|

| Hermes | pip install open-dream-rsi · hermes plugins install "https://github.com/patrykorwat/open-dream-rsi#odr-trigger" · hermes plugins enable odr-trigger — or in the desktop app: Capabilities → Plugins → Install from Git |

| DeepSeek Harness | dsh plugin --profile add dsh-open-dream-rsi (npm bundle — no pip) |

| Claude Code | plugin from this repo: claude plugin marketplace add patrykorwat/open-dream-rsi + claude plugin install open-dream-rsi@open-dream-rsi (prompts for the task/memory paths; no pip) — or pip install open-dream-rsi · claude mcp add open-dream-rsi -- odr mcp --tasks /ABS/PATH/tasks.json --memory /ABS/PATH/.dream_rsi |

| Codex CLI | pip install open-dream-rsi · codex mcp add open-dream-rsi -- odr mcp --tasks /ABS/PATH/tasks.json --memory /ABS/PATH/.dream_rsi |

| OpenCode | pip install open-dream-rsi, then the mcp block in [docs/integrations.md#opencode](docs/integrations.md#opencode) |

| Goose | git clone https://github.com/patrykorwat/open-dream-rsi && cd open-dream-rsi && ./scripts/odr_goose_setup.sh — resolves your provider, starts the proxy, writes the config itself |

| Claude Cowork / claude.ai | remote-only: odr mcp --http --port 8800 behind your tunnel — [docs/integrations.md](docs/integrations.md) |

The odr console script ships with the package (0.3.9+); from a bare

checkout use python3 -m open_dream_rsi or pip install -e ..

Point the dreamer at any OpenAI-compatible endpoint — or nowhere: the model

is auto-detected from /v1/models, so a local vLLM/Ollama needs only the

base URL.

export OPENAI_BASE_URL=http://127.0.0.1:8000/v1   # vLLM / Ollama / any gateway
export OPENAI_API_KEY=***                         # any token for local servers

No agent handy? The bare CLI runs the same loop: `odr --memory .dream_rsi

loop --tasks tasks.json --once (drop --once` for a daemon). No key?

python examples/live_loop_demo.py runs the whole cycle on an in-process

mock server. (dreamrsi on PyPI is an unrelated project — this library is

open-dream-rsi.) Per-host details, MCP blocks, troubleshooting:

[docs/integrations.md](docs/integrations.md). The key-free walkthrough

and the Docker image: [docs/try-it.md](docs/try-it.md).

---

How it works