Any behavioral signal in.
Cognitive architecture out.

Video, transcripts, structured data, behavioral exports — one call returns 9 binary dimensions of cognitive architecture, per-dimension confidence, and an evidence chain. No questionnaire. No cooperation from the subject.

POSTapi.mahakram.in/v1/classify

~3s · async

{
  "subject_id": "sub_4601136C",
  "classification": "Ni/Te BS/C(P)",
  "confidence": "HIGH",
  "evidence": [ /* timestamped clips */ ]
}

Endpoints

Two calls. Classify once, brief forever.

POST/v1/classify

Submit any behavioral signal — video, audio, transcript, structured data (CSV, JSON). Returns a full type code, per-dimension confidence scores, and an evidence chain.

POST/v1/brief

Pass an existing classification + a situation (negotiation, hiring, pitch, managing). Returns a tactical intelligence brief without reclassifying. One classification, unlimited briefs.

Example · classifyPOST /v1/classify

curl https://api.mahakram.in/v1/classify \
  -H "Authorization: Bearer $MAHAKRAM_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "subject_id": "sub_4601136C",
    "sources": [
      { "type": "video", "url": "https://cdn.example.com/interview.mp4" },
      { "type": "csv", "url": "https://cdn.example.com/browsing_data.csv" }
    ]
  }'

# Response
{
  "subject_id": "sub_4601136C",
  "type": "Ni/Te BS/C(P)",
  "confidence": "HIGH",
  "coins": {
    "observer": { "value": "Ni", "p": 0.91 },
    "decider": { "value": "Te", "p": 0.94 },
    "energy_animal": { "value": "Blast", "p": 0.88 },
    "info_animal": { "value": "Sleep", "p": 0.86 }
    // ... 5 more dimensions
  },
  "evidence": [ { "ts": "00:18:42", "coin": "decider", "signal": "..." } ],
  "dossier_url": "https://app.mahakram.in/d/4601136C"
}

Example · briefPOST /v1/brief

curl https://api.mahakram.in/v1/brief \
  -H "Authorization: Bearer $MAHAKRAM_KEY" \
  -d '{
    "subject_id": "sub_4601136C",
    "scenario": "negotiation",
    "context": "Series A term sheet discussion, they want board seat"
  }'

# Response — tactical brief for this specific scenario
{
  "scenario": "negotiation",
  "brief": {
    "approach": "Lead with data, not vision. This architecture...",
    "blind_spots": ["Will over-index on control signals..."],
    "predicted_moves": ["Will push for information rights first..."],
    "what_breaks_the_deal": "Perceived ambiguity in reporting..."
  }
}

Who integrates

Any system that interacts with humans without understanding how they think.

The API accepts any behavioral signal — video, audio, text, structured data exports. It returns cognitive architecture as structured JSON. Every system that personalizes, predicts, or adapts to humans gets better with this layer.

AI agents & personalization

Any system that interacts with humans

Drop the type code into the system prompt. The agent adapts to how this specific person processes information, makes decisions, and responds under pressure. Personalization grounded in cognition, not click history.

Consumer AI · Voice agents · Coaching apps · AI companions · Recommendation engines

Talent infrastructure

Classify from any recorded interview

No questionnaire required from the candidate. Every interview returns cognitive architecture, predicted ceilings, failure modes, and per-role fit — all derived from observation.

ATS vendors · Interview platforms · Exec search · Talent intelligence

Enterprise intelligence

A cognitive layer on every contact record

The buyer's cognitive type attached to the CRM record. Predicted objections before the call. The framing that lands for this specific architecture. The move that loses the deal.

CRM · Sales enablement · Deal intelligence · Negotiation coaching

Behavioral data platforms

You already have the signal — we add the interpretation

Browsing patterns, purchase records, content consumption, app usage, structured exports. We turn raw behavioral data into cognitive architecture without needing a single recording or transcript.

Consumer analytics · AdTech · Recommendation systems · Data enrichment

Early access.
One pilot running. Onboarding the next.

Tell us what data you’re sitting on and what a cognitive layer would unlock. We’ll tell you if the current pipeline supports your input types.

Email abhas@mahakram.in →Request access form

Any behavioral signal in.Cognitive architecture out.