Prism API - Multidisciplinary Problem Analysis
What is Prism?
Prism is an API that analyzes a given problem through the principles of different academic disciplines. Just as a prism splits white light into its components, Prism splits a problem into its disciplinary perspectives.
The system uses deepseek-v4-flash on the backend. For any given problem, it selects the most interesting and unexpected disciplines, then reinterprets the problem from each discipline's framework.
How It Works
You make a single call. Send the problem text, and Prism returns:
- The problem's primary domain, sub-domain, and type
- The selected disciplines and why each is interesting
- A full perspective from each discipline: interpretation, approach, key insight, and alternative solution
All of this happens in a single LLM call, making it fast and cost-effective.
Modes
Prism offers three modes:
- intra: Stays within the problem's own broad field, picking sub-disciplines with different methods. For example, approaching an algebra problem through geometry or number theory.
- inter: Picks disciplines from entirely different fields. For example, approaching an algebra problem through anthropology or mythology.
- both: Mixes both same-field and cross-field perspectives.
API Reference
Base URL
https://multi.niyazi.cv
POST /analyze
The main endpoint. Analyzes a problem.
Headers:
Content-Type: application/json
X-API-Key: ***
Request Body:
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
| problem | string | Yes | - | The problem to analyze (max 5000 chars) |
| mode | string | No | both | intra, inter, or both |
| disciplines | string or array | No | auto | auto or a list of discipline names |
| count | number | No | 2 | Number of perspectives (1-8) |
| depth | string | No | brief | brief or detailed |
Example Request:
curl -X POST https://multi.niyazi.cv/analyze \
-H "Content-Type: application/json" \
-H "X-API-Key: ***" \
-d '{
"problem": "A team of 5 developers keeps missing deadlines. They work long hours but deliver buggy code.",
"mode": "both",
"count": 3,
"depth": "brief"
}'
Response (200 OK):
{
"domain": "Software Development Management",
"sub_domain": "Team Productivity",
"problem_type": "Chronic schedule slippage",
"perspectives": [
{
"discipline": "Software Engineering",
"why": "Directly addresses technical causes of buggy code.",
"mode": "intra",
"interpretation": "The team is accumulating technical debt...",
"approach": "1. Conduct post-mortem. 2. Implement CI/CD. 3. Adopt agile estimation.",
"key_insight": "Long hours amplify technical debt.",
"alternative_solution": "Institute a quality gate with automated tests."
}
]
}
GET /health
Service health check.
Response:
{
"status": "ok",
"model": "deepseek-v4-flash:0731-cloud"
}
Discipline Selection
Prism delegates discipline selection to the LLM. This is by design: the goal is creative and unexpected choices.
In auto mode, the model picks the most interesting disciplines based on the problem's nature. Examples:
- Math problem gets Mathematical Logic or Cognitive Psychology
- Software problem gets Epidemiology or Music Theory
- Economics problem gets Paleontology or Cultural Anthropology
Manual selection is also supported. Pass a list of disciplines in the disciplines parameter, and Prism will generate perspectives from those specific fields.
Use Cases
Education
Show students multiple approaches to the same math problem through geometry, number theory, and statistics, enriching their thinking.
Research
Discover cross-disciplinary connections. Approach a biology problem through chemistry or physics to find new angles.
Product and UX Design
Analyze design problems from psychology, anthropology, and economics perspectives.
Strategy
Examine business challenges through biology, game theory, and history.
Content Creation
Generate unique perspectives for essays and articles.
Pricing
Prism API is available on RapidAPI.
| Plan | Price | Requests/day | Max perspectives | Features |
|---|---|---|---|---|
| Basic | Free | 1 | 2 | Brief mode |
| Pro | $5/mo | 5 | 3 | Brief + detailed |
| Ultra | $9/mo | 10 | 5 | All features |
| Mega | $19/mo | 25 | 8 | All features + priority |
A call to any endpoint counts as one request.
Technical Details
- Backend: Python FastAPI
- LLM: deepseek-v4-flash (Ollama Cloud)
- Host: multi.niyazi.cv
- Auth: X-API-Key header
- Average Response Time: 6-12 seconds
- Max Problem Length: 5000 characters
Quick Start
import requests
response = requests.post(
"https://multi.niyazi.cv/analyze",
headers={
"Content-Type": "application/json",
"X-API-Key": "***"
},
json={
"problem": "How can I reduce my carbon footprint while traveling?",
"mode": "both",
"count": 3,
"depth": "brief"
}
)
data = response.json()
for p in data["perspectives"]:
print(f"\n--- {p['discipline']} ---")
print(f"Why: {p['why']}")
print(f"Insight: {p['key_insight']}")
print(f"Solution: {p['alternative_solution']}")