ARISE — Adaptive Runtime Improvement through Self-Evolution
Your agent works great on the tasks you planned for. ARISE handles the ones you didn't.
ARISE is a framework-agnostic middleware that gives LLM agents the ability to create their own tools at runtime. When your agent fails at a task, ARISE detects the capability gap, synthesizes a Python tool, validates it in a sandbox, and promotes it to the active library — no human intervention required.
Documentation | Quick Start | PyPI
pip install arise-ai
from arise import ARISE
from arise.rewards import task_success
arise = ARISE(
agent_fn=my_agent, # any (task, tools) -> str function
reward_fn=task_success,
model="gpt-4o-mini", # cheap model for tool synthesis
)
result = arise.run("Fetch all users from the paginated API")
# Agent fails → ARISE synthesizes fetch_all_paginated tool → agent succeeds
What It Looks Like
Episode 1 | FAIL | reward=0.00 | skills=2 Task: "Fetch paginated users with auth"
Episode 2 | FAIL | reward=0.00 | skills=2
Episode 3 | FAIL | reward=0.00 | skills=2
[Evolution triggered — 3 failures on API tasks]
→ Synthesizing 'parse_json_response'... 3/3 tests passed ✓
→ Synthesizing 'fetch_all_paginated'... sandbox fail → refine → 1/1 passed ✓
Episode 4 | OK | reward=1.00 | skills=4 Agent now has the tools it needs
Key Features
- Self-evolving tool library — fail → detect gap → synthesize → sandbox test → promote
- Framework-agnostic — any
(task, tools) -> strfunction, Strands, LangGraph, CrewAI - Sandboxed validation — subprocess or Docker, adversarial testing, import restrictions
- Distributed mode — S3 + SQS for stateless deployments (Lambda, ECS, AgentCore)
- Skill registry — share evolved tools across projects
- Version control + rollback — SQLite checkpoints,
arise rollback <version> - A/B testing — refined skills tested against originals before promotion
- Web Console — create agents, watch evolution live, inspect evolved code (
arise console) - Dashboard — terminal TUI and web UI for monitoring
Benchmark Results
| Model | Condition | AcmeCorp (SRE) | DataCorp (Data Eng) |
|---|---|---|---|
| Claude Sonnet | ARISE | 78% | — |
| Claude Sonnet | No tools | 63% | — |
| GPT-4o-mini | ARISE | 57% | 92% |
| GPT-4o-mini | No tools | 48% | 50% |
ARISE improves task success by +9–42 percentage points across models and domains. See the full benchmark results.
ARISE Console
A web UI for creating agents, watching evolution live, and inspecting evolved tools:
arise console
# Opens http://localhost:8080
- Create agents — pick model, set system prompt, choose reward function
- Live terminal feed — watch episodes and evolution in real-time via WebSocket
- Skill inspector — syntax-highlighted code, test suite, performance metrics
- Editable config — change reward function, system prompt, failure threshold on the fly
- All Skills / Evolution Log — global views across all agents
Documentation
Full documentation at arise-ai.dev:
- Installation — install and configure
- Quick Start — complete evolution loop walkthrough
- How It Works — the 5-step evolution pipeline
- Reward Functions — built-in and custom reward functions
- Safety & Validation — sandbox, adversarial testing, production recommendations
- Distributed Mode — S3 + SQS for stateless deployments
- Framework Adapters — Strands, LangGraph, CrewAI, raw OpenAI/Anthropic
- CLI Reference — all CLI commands
- API Reference — ARISE class, config, types
Examples
| Example | Description |
|---|---|
quickstart_evolution.py |
Full evolution loop: agent fails → ARISE evolves tool → agent succeeds |
quickstart.py |
Math agent evolves statistics tools |
api_agent.py |
HTTP agent evolves auth + pagination (mock server) |
devops_agent.py |
DevOps agent evolves log analysis tools |
strands_agent.py |
Strands integration with Bedrock |
demo/agentcore/ |
AgentCore deployment with A2A protocol |
Install
pip install arise-ai # core (just pydantic)
pip install arise-ai[aws] # + boto3 for distributed mode
pip install arise-ai[litellm] # + litellm for multi-provider LLM
pip install arise-ai[docker] # + docker sandbox backend
pip install arise-ai[dashboard] # + rich, fastapi for dashboard
pip install arise-ai[otel] # + opentelemetry for tracing
pip install arise-ai[all] # everything
Related Work
ARISE builds on ideas from LATM, VOYAGER, CREATOR, ADAS, and CRAFT. ARISE adds the production layer: framework-agnostic integration, sandboxed validation, adversarial testing, version control, distributed deployment, and A/B testing.
License
MIT
Metadata
Release files for arise-ai 0.2.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| arise_ai-0.2.1-py3-none-any.whl | Python 3 | none | any | Details for arise_ai-0.2.1-py3-none-any.whl |
Release files / arise_ai-0.2.1-py3-none-any.whl
| Download URL | arise_ai-0.2.1-py3-none-any.whl |
|---|---|
| Size | 79.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
2f95dc834ef004c461a69d54afd32466c9333bf785879bc7a3c5c699d85809ee
|
|
BLAKE2b-256 checksum How to use checksums |
9f9f529d29561900342bbcc915a4eea288cd7277458cde33488d8cd1b5dfcaef
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.4
|
Security for arise-ai 0.2.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.