Validate memory persistence and routing with Telegram and Qdrant

Define the household test scenario

You’ll create a shared household assistant to test local memory. You’ll save and retrieve a synthetic maintenance reminder without sending it to a public cloud LLM.

Telegram transports the messages. Ollama, Qdrant, and the local LLM process them on your host.

Store and query local memory

Send the following command to the Telegram bot:

    

        
        
/mem #home The boiler should be inspected every October.

    

The runtime stores the reminder through the following path:

    

        
        
Telegram / Mem command
    -> Memory skill
    -> Ollama embedding
    -> Qdrant collection: personal_tracker_memory

    

Wait for the confirmation, then retrieve the memory:

    

        
        
/rag memory: When should the boiler be inspected?

    

The response should mention October:

Saving and retrieving a household memory in Telegram

The retrieval request uses the following local path:

    

        
        
Telegram question
    -> Ollama query embedding
    -> Qdrant similarity search
    -> Retrieved context
    -> Local vLLM response
    -> Telegram answer

    

Verify Qdrant vector collections

Confirm that the personal memory collection exists:

    

        
        
curl http://127.0.0.1:6333/collections/personal_tracker_memory

    

The relevant fields are similar to:

    

        
        {
  "result": {
    "status": "green",
    "optimizer_status": "ok",
    "points_count": 102,
    "config": {
      "params": {
        "vectors": {
          "size": 768,
          "distance": "Cosine"
        },
        "on_disk_payload": true
      }
    }
  },
  "status": "ok"
}

        
    

The point count depends on existing data. A green status with optimizer_status set to ok confirms collection health. The vector size of 768 matches nomic-embed-text.

The collection metadata doesn’t prove that the boiler reminder was stored. Query the point payload directly to verify the synthetic record:

    

        
        
curl -sS -X POST \
  http://127.0.0.1:6333/collections/personal_tracker_memory/points/scroll \
  -H 'Content-Type: application/json' \
  -d '{
    "filter": {
      "must": [
        {
          "key": "text",
          "match": {
            "value": "#home The boiler should be inspected every October."
          }
        }
      ]
    },
    "limit": 5,
    "with_payload": true,
    "with_vector": false
  }'

    

Look for the boiler reminder in the returned payload. The filter finds it even when the personal collection contains other records. This verifies the stored data directly instead of relying on the assistant’s response.

Inspect active agents and task execution

Send the following command to the Telegram bot:

    

        
        
/agents

    

The response lists the thin agents registered by the reference runtime, including memory, RAG, browser search, weather, and chat routes.

To inspect recent tasks, send the following command to the Telegram bot:

    

        
        
/tasks last 5

    

Task history shows which agent handled the request, its status, and its runtime. All routes use the configured LLM endpoint.

Test external skill integration

Send a weather question in plain language:

    

        
        
Cambridge weather tomorrow

    

The runtime sends this question to the weather skill. Don’t add /search, which selects the general browser worker instead.

This request contacts the public wttr.in weather service, but generation still uses the local model.

Your household assistant should now:

1. Save and retrieve the synthetic boiler reminder from Telegram.
2. Store the reminder in personal_tracker_memory.
3. Show the selected agent in /agents and /tasks last 5.
4. Return weather data through the external weather skill.

What you’ve accomplished and what’s next

You’ve now saved and retrieved a synthetic household memory, verified it in Qdrant, and inspected both local and external request paths.

Next, you’ll add document RAG, browser search, and a proactive cron reminder.