Anthropic

4 production-ready jobs — long document analysis for persona creation, transcript mega-analysis for quarterly intelligence, multi-source synthesis for brand voice extraction, and research report focus group question generation

Anthropic builds Claude, a family of frontier AI models with industry-leading context windows (up to 1M tokens). These four jobs leverage Claude’s ability to process massive documents in a single call — entire annual reports, quarters of meeting transcripts, multi-source writing samples, and full research papers — then feed the extracted intelligence into Mavera’s surfaces (Personas, Mave Agent, Brand Voices, Focus Groups) to create data-driven personas, enriched intelligence reports, authentic brand voices, and validated research questions.

---

API Reference Card

Detail	Value
Base URL	https://api.anthropic.com
Auth	x-api-key: {key} header (NOT Bearer)
Required header	anthropic-version: 2023-06-01
Models	claude-opus-4-6-20250725 (1M tokens), claude-sonnet-4-20250514 (200K), claude-haiku-3-5-20241022 (200K)
Rate limits	Tier-dependent — Tier 1: 50 req/min, Tier 4: 4,000 req/min
Mavera base	https://app.mavera.io/api/v1
Mavera auth	Authorization: Bearer mvra_live_...

All examples use ANTHROPIC_API_KEY for Claude and MAVERA_API_KEY for Mavera. Anthropic uses x-api-key — not Authorization: Bearer. Every request also requires the anthropic-version header. The Python SDK and JS SDK handle both headers automatically.

---

Prerequisites

1. Anthropic API key

   Create an account at console.anthropic.com and generate an API key from the Settings page.
2. Mavera API key

   Get your key from Mavera dashboard.
3. Install SDKs

   pip install anthropic requests
   

   or for JavaScript:

   npm install @anthropic-ai/sdk
   
4. Environment variables

   export ANTHROPIC_API_KEY="sk-ant-api03-xxxxx"
   export MAVERA_API_KEY="mvra_live_xxxxx"
   

---

Jobs

#	Job	Claude Capability	Mavera Surface	Output
1	Long Document Analysis → Persona Source	1M token context (Opus 4.6)	Personas	Data-driven personas from full documents
2	Transcript Mega-Analysis	Long context cross-analysis	Mave Agent (Chat)	Enriched quarterly intelligence report
3	Multi-Source Synthesis → Brand Voice	Pattern extraction across sources	Brand Voices	Authentic brand voice profile
4	Research Report → Focus Group Questions	Finding extraction + synthesis	Personas + Focus Groups	Validated research with auto-generated questions

---

Rate Limits & Production Notes

Model	Tier 1	Tier 4	Cost (Input / Output per M tokens)	Strategy
Claude Opus 4.6	50 req/min	4,000 req/min	$15 /$75	Reserve for documents >100K tokens
Claude Sonnet	50 req/min	4,000 req/min	$3 /$15	Default for most analysis tasks
Claude Haiku	50 req/min	4,000 req/min	$0.25 /$1.25	Use for extraction and formatting

Anthropic uses x-api-key authentication — not Authorization: Bearer. Every request requires the anthropic-version: 2023-06-01 header. The Python and JS SDKs handle both automatically. For raw HTTP calls, always include:

headers = {"x-api-key": key, "anthropic-version": "2023-06-01", "Content-Type": "application/json"}

Large context calls (500K+ tokens) can take 60-120 seconds — set timeouts accordingly. Cache document analyses to avoid reprocessing unchanged content. Monitor Mavera credits at Dashboard.

For cost optimization, use a two-model strategy: Claude Haiku for initial extraction and formatting ($0.25/M tokens), then Claude Opus 4.6 only for deep analysis requiring reasoning across full context ($15/M tokens). This reduces costs by up to 80% on mixed workloads.

---

All Integrations

Anthropic API Docs

Personas

Focus Groups

Brand Voices

Mave Agent