Integration

OpenClaw Memory & Activity Tracking

Give your OpenClaw AI agent persistent memory that survives restarts and a full activity audit trail — powered by OpenSink.

The Problem

OpenClaw agents store memory in local files — but that memory is tied to one machine, invisible to dashboards, and there's no audit trail of what your agent actually did.

Amnesia Between Sessions

Your OpenClaw agent wakes up fresh every session. Past conversations, decisions, user preferences — all gone. It asks the same questions twice.

No Audit Trail

Your OpenClaw agent ran a cron job at 3am. What did it do? Who did it message? Without structured logging, you're guessing.

No Observability

Your agent handles Telegram messages, runs sub-agents, processes tasks — but you have no dashboard, no timeline, no way to inspect what happened.

Two Skills, Full Control

Install opensink-memory for persistent agent memory and opensink-skills for activity logging and session management. Both are open source, both install in minutes.

∞

retention

Persistent Memory

Install the opensink-memory skill and your OpenClaw agent remembers facts, preferences, and decisions across sessions. It learns who you are over time.

100%

visibility

Full Activity Timeline

Install the opensink-skills pack and every significant action — emails sent, reports generated, decisions made — is logged as an Activity. View the full timeline in your OpenSink dashboard.

<5min

setup time

Two Skills, Five Minutes

Install opensink-memory and opensink-skills from ClawHub or GitHub. Add your API key. That's it — your agent handles the rest, creating its own sessions and memory sink automatically.

Real-time

dashboard

Dashboard Included

Everything your agent remembers and everything it does is visible in the OpenSink dashboard. Search memories, inspect activity timelines, audit sessions — all in one place.

Install, Configure, Done

Store a Memory

memory.sh

# Install the memory skill
clawhub install opensink-memory

# Set your API key — that's the only config needed
export OPENSINK_API_KEY="osk_..."

# Your agent stores memories automatically
bash scripts/opensink-memory.sh push "Dan prefers dark mode" "preference"
bash scripts/opensink-memory.sh push "Deploy on Fridays = bad" "decision"

# Search across all memories
bash scripts/opensink-memory.sh search "dark mode"
# → [preference] Dan prefers dark mode (2026-02-09)

Log Activities

activity.sh

# Install the skills pack
clawhub install opensink-skills

# Create an agent (first time only)
AGENT_ID=$(bash scripts/agent.sh create "Mars")

# Start a session — tracks everything your agent does
SESSION_ID=$(bash scripts/session.sh start)

# Log activities as your agent works
bash scripts/activity.sh log "Processed 12 emails" "message"
bash scripts/activity.sh log "Sent weekly report" "message" \
  '{"emails": 12, "flagged": 3}'
bash scripts/activity.sh log "Escalated ticket #42" "message"

# Mark session complete
bash scripts/session.sh status $SESSION_ID completed
# → Full timeline visible in OpenSink dashboard

Before vs After OpenSink

Feature	Without OpenSink	With OpenSink
Memory persistence	Gone on restart	opensink-memory skill — permanent, searchable
Activity tracking	Parse logs manually	opensink-skills — structured timeline + dashboard
Session management	No structure	opensink-skills — auto-created, tracked sessions
Setup	Build your own infrastructure	Two skills + one API key, 5 minutes
Cross-session context	Start from scratch every time	Agent remembers everything automatically
Observability	Blind	Full dashboard — memories, activities, sessions

Give your agent a memory

Two open-source skills. One API key. Five minutes to set up.

Get Started View on GitHub