Dumb Agent

The dumb agent is a stateless wrapper around the core writing system. You send it text, it sends back a result. No memory. No chaining. No session state.

Use this when you want to integrate Anky's writing analysis into your own system without any magic.

Endpoints

Submit Writing

bash

POST /agents/dumb/write

Submit raw writing text and get back analysis + feedback.

Headers:

Authorization: Bearer <session_token>
Content-Type: application/json

Request:

json

{
  "writing": "your stream of consciousness text here...",
  "duration_seconds": 512,
  "new_line_count": 3
}

Response (under threshold):

json

{
  "is_anky": false,
  "feedback": "You wrote for 8 minutes but the flow score was low. There were 47 pauses longer than 3 seconds. The writing feels like you were thinking about what to write rather than writing what you were thinking. Try again with your eyes closed.",
  "flow_score": 0.34,
  "duration_seconds": 512,
  "threshold_met": false
}

Response (real anky):

json

{
  "is_anky": true,
  "writing_id": "550e8400-e29b-41d4-a716-446655440000",
  "flow_score": 0.87,
  "duration_seconds": 512,
  "chakra": 4,
  "kingdom": "Eleasis",
  "feedback": null
}

Get Writing Prompt

bash

GET /agents/dumb/prompt

Get a random writing prompt. No personalization. No memory. Just a starting point.

Response:

json

{
  "prompt": "Write about a door you chose not to open.",
  "source": "ankyverse:eleasis"
}

Analyze Text

bash

POST /agents/dumb/analyze

Analyze a piece of text without persisting it. Returns chakra resonance and thematic breakdown without triggering any pipelines.

Request:

json

{
  "text": "any text to analyze..."
}

Response:

json

{
  "chakra": 5,
  "kingdom": "Voxlumis",
  "themes": ["expression", "silence", "truth-telling"],
  "emotional_signature": "restrained urgency",
  "flow_estimate": 0.72
}

Usage Example

python

import requests

API = "https://anky.app/agents/dumb"
TOKEN = "your-session-token"
headers = {"Authorization": f"Bearer {TOKEN}"}

# Get a prompt
prompt = requests.get(f"{API}/prompt", headers=headers).json()
print(f"Write about: {prompt['prompt']}")

# Submit writing
result = requests.post(f"{API}/write", headers=headers, json={
    "writing": "eight minutes of raw text...",
    "duration_seconds": 490,
    "new_line_count": 2
}).json()

if result["is_anky"]:
    print(f"Anky created in {result['kingdom']}")
else:
    print(f"Feedback: {result['feedback']}")

When to Use Dumb

You're building your own writing app and just want the analysis
You want to batch-analyze existing texts for chakra mapping
You're testing the system before committing to a session
You want full control over the UX and don't want the agent making decisions