agent-memory

# Agent Memory — Telegram Memory for AI Agents Gives you persistent long-term memory across Telegram channels and groups. Search conversations, generate digests, extract decisions and action items. ## First Run (if AGENT_MEMORY_API_KEY is not set) When the user asks anything related to Telegram memory, channels, or this skill, and `AGENT_MEMORY_API_KEY` is not configured, walk them through setup: 1. Tell the user: "To use Telegram memory, you need an API key. Open @AgentMemoryBot in Telegram: https://t.me/AgentMemoryBot" 2. In the bot: press /start → 📱 Connect Telegram (share contact, enter reversed code) → 🔑 API Keys → ➕ Create Key 3. Ask the user to paste the key (starts with `amk_`) 4. Once they give you the key, save it to the OpenClaw config: ```bash # Add to openclaw.json skills config cat ~/.openclaw/openclaw.json | jq '.skills.entries["agent-memory"] = {"env": {"AGENT_MEMORY_API_KEY": "USER_KEY_HERE"}}' > /tmp/oc.json && mv /tmp/oc.json ~/.openclaw/openclaw.json ``` Or if openclaw.json doesn't exist or has no skills section, create it. Then tell the user to restart the OpenClaw session. ## How to Use All API calls use: `Authorization: Bearer $AGENT_MEMORY_API_KEY` Base URL: `https://agent.ai-vfx.com` ### Search Memory Find information across synced Telegram channels. Use for any question about channel content. ```bash curl -s -X POST https://agent.ai-vfx.com/api/v1/memory/search \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $AGENT_MEMORY_API_KEY" \ -d '{"query": "what was discussed about topic X", "scope": "@channel", "limit": 10}' ``` - `scope`: optional — `@username` for one channel, `folder:Name` for a folder, omit for all - Response: `{"answer": "...", "sources": [{"url": "https://t.me/...", "channel": "..."}]}` ### List Sources See which channels are connected. ```bash curl -s https://agent.ai-vfx.com/api/v1/sources \ -H "Authorization: Bearer $AGENT_MEMORY_API_KEY" ``` ### Add Source Connect a new Telegram channel. User must have connected their Telegram in the bot first. ```bash curl -s -X POST https://agent.ai-vfx.com/api/v1/sources/add \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $AGENT_MEMORY_API_KEY" \ -d '{"handle": "@channel_username", "sync_range": "1m"}' ``` sync_range options: 1w, 1m, 3m, 6m, 1y ### Get Digest Summarize conversations for a period. ```bash curl -s -X POST https://agent.ai-vfx.com/api/v1/digest \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $AGENT_MEMORY_API_KEY" \ -d '{"scope": "@channel", "period": "7d"}' ``` period options: 1d, 3d, 7d, 30d ### Extract Decisions Get decisions, action items, and open questions. ```bash curl -s -X POST https://agent.ai-vfx.com/api/v1/decisions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer $AGENT_MEMORY_API_KEY" \ -d '{"scope": "@channel", "topic": "optional filter"}' ``` ### Sync Status Check if channels are still syncing after add_source. ```bash curl -s https://agent.ai-vfx.com/api/v1/sync-status \ -H "Authorization: Bearer $AGENT_MEMORY_API_KEY" ``` ### Balance Check remaining points. ```bash curl -s https://agent.ai-vfx.com/api/v1/account/balance \ -H "Authorization: Bearer $AGENT_MEMORY_API_KEY" ``` ## When to Use Which Operation - "What did they discuss about X?" → **search** with the topic as query - "What happened this week in @channel?" → **digest** with period=7d - "What decisions were made?" → **decisions** - "Connect @newchannel" → **add_source**, then wait and check **sync_status** - "What channels do I have?" → **list_sources** - "How many points left?" → **balance** ## Costs 1 point ≈ $0.01. Search: 3 pts, Digest: 25 pts, Decisions: 12 pts. Top up via TON in @AgentMemoryBot.