🧶 StateWeave

StateWeave is git for agent brains — debug, time-travel, and migrate agent state across 10 frameworks. When a 20-step autonomous workflow derails at step 15, see exactly what changed, rewind to step 14, and replay. Export from LangGraph, import into CrewAI with zero data loss. Checkpoint, rollback, diff, encrypt, sign — all through a single Universal Schema.

When your agent hallucinates, crashes, or drifts — stateweave why shows you the exact state transition that went wrong. When your enterprise needs to audit agent behavior, every state change is versioned, signed, and encrypted.

Why StateWeave?

StateWeave solves three critical problems in the AI agent ecosystem:

🔍 Debugging — Agent workflows are non-deterministic. When they go wrong, you need to pause, rewind, inspect, and replay — not restart. stateweave why shows you the exact state transition that caused the failure. Version control for agent cognition.

🔒 Security — Agent state contains the agent's entire cognitive history. StateWeave encrypts at rest (AES-256-GCM), signs payloads (Ed25519), strips credentials on export, and enforces compliance policies.

🔄 Portability — Every framework has persistence, none have portability. StateWeave's Universal Schema — a canonical representation of agent cognitive state — lets you move state between any of 10 frameworks. One schema, N adapters, zero data loss (with explicit warnings for anything non-portable).

┌─────────────┐     ┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│  LangGraph  │     │    MCP      │     │   CrewAI    │     │   AutoGen   │
│   Adapter   │     │   Adapter   │     │   Adapter   │     │   Adapter   │
└──────┬──────┘     └──────┬──────┘     └──────┬──────┘     └──────┬──────┘
       │                   │                   │                   │
       └───────────┬───────┴───────────┬───────┘                   │
                   │                   │                           │
                   ▼                   ▼                           ▼
            ┌──────────────────────────────────────────────────────────┐
            │              🧶  Universal Schema v1                     │
            │                                                          │
            │  conversation_history  ·  working_memory  ·  goal_tree   │
            │  tool_results_cache  ·  trust_parameters  ·  audit_trail │
            └──────────────────────────────────────────────────────────┘

Star topology, not mesh. N adapters, not N² translation pairs. Adding a new framework = one adapter, instant compatibility with everything else.

See it working

$ pip install stateweave
$ python examples/full_demo.py

━━ 1. Export from LangGraph ━━
  ✓ Exported 4 messages
  ✓ Source framework: langgraph

━━ 2. Import into MCP ━━
  ✓ Imported into mcp
  ✓ Messages preserved: 4

━━ 3. Verify Round-Trip ━━
  ✓ Zero data loss: YES

━━ 4. Diff Agent States ━━
  Summary: 7 added, 4 removed, 7 modified

━━ 5. Time Travel ━━
  ✓ Checkpoint v1 (initial-research)
  ✓ Checkpoint v2 (after-drug-discovery)
  ✓ Rolled back → 4 msgs

━━ 6. Encryption (AES-256-GCM) ━━
  ✓ 1,733 bytes → 1,749 bytes encrypted
  ✓ Decrypted: 4 messages intact

━━ 7. Non-Portable Warnings ━━
  ✓ No non-portable warnings (clean export)

7/7 steps passed. Everything runs from PyPI.

Try it now: pip install stateweave && stateweave quickstart — zero-code demo in 10 seconds.

Or run the full 7-step demo: python examples/full_demo.py

One-Command Migration

$ stateweave migrate --from langgraph --to crewai --agent my-agent

  🧶 StateWeave Migrate: langgraph → crewai
  ════════════════════════════════════════════════

  ━━ Step 1: Export from langgraph ━━
    ✓ Exported 12 messages, 5 memory keys (0.01s)

  ━━ Step 2: Validate payload ━━
    ✓ Payload valid — all schema checks passed

  ━━ Step 3: Import into crewa