palaia 2.2.3

Local, cloud-free memory for OpenClaw agents.

Palaia v2.2 — The Knowledge OS for OpenClaw Agent Teams

Crash-safe. Local-first. Zero-cloud. The memory system that makes your agents smarter over time.

---

Install

Recommended: Tell your agent

Paste this into your OpenClaw agent:

Install palaia for me. Set up memory so it works across sessions.

The agent handles everything: pip install "palaia[fastembed]", palaia init, plugin setup, and verification.

Manual / Expert Setup

pip install "palaia[fastembed]"
palaia init
palaia doctor --fix

For the OpenClaw plugin (Auto-Capture + Auto-Recall):

npm install -g @byte5ai/palaia@latest

Then configure openclaw.json — see SKILL.md for details.

For knowledge curation features:

pip install "palaia[curate]"

Upgrading? pip install --upgrade "palaia[fastembed]" && palaia doctor --fix — migration is automatic.

Quick Start

palaia init                                         # Initialize (SQLite store)
palaia write "API rate limit is 100 req/min" \
  --type memory --tags api,limits                   # Save knowledge
palaia query "what's the rate limit"                # Find it by meaning

---

What's New in v2.2

• SQLite as default backend — Zero-config, single-file database with WAL mode. Replaces flat JSON files. Existing stores migrate automatically.
• Injection priorities — Per-agent/project control over which memories get injected (palaia priorities).
• Knowledge curation — palaia curate analyze/apply for clustering, dedup, and clean migration.
• ContextEngine integration — New OpenClaw ContextEngine adapter with 7 lifecycle hooks.
• Service layer — Business logic extracted into palaia/services/ package.
• Doctor decomposition — palaia/doctor/ package with modular checks, fixes, detection.
• New nudges — Contextual guidance for curation, priorities, and backend migration.

See CHANGELOG.md for full details.

---

Why Palaia?

Feature	Details
SQLite-Backed Storage	Single-file database with WAL mode. Zero dependencies (sqlite3 is stdlib). Optional PostgreSQL for distributed teams.
Semantic Search	Hybrid BM25 + embeddings. Providers: fastembed, sentence-transformers, OpenAI, Gemini, Ollama
Crash-Safe Writes	WAL-backed — survives power loss, kills, OOM
Auto-Capture	OpenClaw plugin captures significant exchanges automatically
Structured Types	memory, process, task — with status, priority, assignee fields
Multi-Agent	Shared store, scopes (private/team/public), agent aliases, per-agent priorities
Smart Tiering	HOT -> WARM -> COLD rotation based on access patterns
Injection Priorities	Per-agent/project control over what gets injected into context
Knowledge Curation	Cluster, deduplicate, and clean up accumulated knowledge for migration
Adaptive Nudging	Teaches agents best practices, graduates when they learn
Zero-Cloud	Everything runs locally. No API keys required for core functionality

---

Architecture

palaia/
  backends/        Storage backends (SQLite, PostgreSQL)
  services/        Business logic (write, query, status, admin)
  doctor/          Diagnostics (checks, fixes, detection)
  hooks/           OpenClaw hook handlers (recall, capture, state, reactions)
  context-engine   ContextEngine adapter (7 lifecycle hooks)

See ARCHITECTURE.md for the full module map and data flows.

---

CLI Reference

palaia init                                          Initialize store (SQLite)
palaia write "text" [--type TYPE] [--tags a,b]       Save knowledge
palaia query "search" [--type TYPE] [--project P]    Search by meaning
palaia get <id>                                       Read specific entry
palaia list [--tier T] [--type T] [--status S]       List entries
palaia edit <id> [--status done]                      Edit entry
palaia status                                         System health
palaia doctor [--fix]                                 Diagnose + fix
palaia project create|list|show|query|delete          Manage projects
palaia memo send|inbox|ack|broadcast                  Inter-agent messaging
palaia priorities [block|set]                         Injection priorities
palaia curate analyze|apply                           Knowledge curation
palaia sync export|import                             Git-based exchange
palaia package export|import|info                     Portable packages
palaia process list|run                               Process tracking
palaia gc [--aggressive] [--budget N]                 Garbage collection
palaia config list|set|set-chain                      Configuration
palaia ingest <source> [--project P]                  Index documents (RAG)
palaia detect                                         Available providers
palaia warmup                                         Pre-build search index

All commands support --json for machine-readable output.

---

Comparison

Feature	Palaia	Stock Memory	Mem0	Engram
Local-first	Yes	Yes	No (cloud)	Yes
Crash-safe (WAL)	Yes	No	N/A	No
SQLite Backend	Yes (default)	No	No	No
Auto-Capture	Yes (plugin)	No	Yes	No
Structured Types	Yes (memory/process/task)	No	No	No
Multi-Agent Scopes	Yes (private/team/public)	No	Per-user	No
Per-Agent Priorities	Yes	No	No	No
Smart Tiering	Yes (HOT/WARM/COLD)	No	No	No
Knowledge Curation	Yes	No	No	No
Semantic Search	Hybrid (embedding + BM25)	None	Embedding	Embedding
Zero-Cloud	Yes	Yes	No	Yes

---

Development

git clone https://github.com/iret77/palaia.git
cd palaia
pip install -e ".[dev]"
pytest

Links

• GitHub — Source + Issues
• PyPI — Package registry
• ClawHub — Install via agent skill
• OpenClaw — The agent platform Palaia is built for
• CHANGELOG — Release history

---

MIT — (c) 2026 byte5 GmbH