OpenAI - Mavera

OpenAI’s model suite — GPT-5, GPT-4.1, o3/o4-mini for reasoning, Whisper for speech-to-text, TTS for text-to-speech, DALL-E and Sora for visual generation, and embedding models for vector search — pairs with Mavera to create multi-modal content pipelines. These five jobs chain OpenAI endpoints with Mavera surfaces to transcribe meetings into structured action plans, pre-process video with vision models, convert generated copy into audio libraries, find and fill content gaps via embeddings, and run video generation-analysis feedback loops.

API Reference Card

Detail	Value
Base URL	`https://api.openai.com/v1`
Auth	Bearer Token — `Authorization: Bearer sk-...`
Key models	GPT-5, GPT-4.1, o3/o4-mini, Whisper, TTS, DALL-E 3, text-embedding-3-large, Sora
Rate limits	Vary by model tier — TPM and RPM limits per organization (see OpenAI rate limits)
Mavera base	`https://app.mavera.io/api/v1`
Mavera auth	`Authorization: Bearer mvra_live_...`

All examples use two environment variables: OPENAI_API_KEY (your OpenAI platform key starting with sk-) and MAVERA_API_KEY (your Mavera key starting with mvra_live_). Never commit either key to version control. Use a .env file or your platform’s secret manager.

Prerequisites

OpenAI API key

Sign up at platform.openai.com. Navigate to API keys and create a new secret key. Ensure your organization has billing enabled and sufficient credits.

Mavera API key

Get your key from Mavera dashboard.

Install SDKs

# Python
pip install openai requests

# JavaScript
npm install openai

Set environment variables

export OPENAI_API_KEY="sk-your-openai-key"
export MAVERA_API_KEY="mvra_live_xxxxx"

Jobs

#	Job	OpenAI Endpoint	Mavera Surface	Output
1	Whisper Transcription → Meetings Pipeline	POST /audio/transcriptions	Mave Agent (Chat)	Structured meeting analysis
2	GPT Vision → Video Analysis Pre-Processing	Responses (vision)	Mave Agent (Chat)	Marketing analysis of visual content
3	TTS → Audio Content Library	POST /audio/speech	Generate	Audio files from generated content
4	Embeddings → Knowledge Base Gap Filling	POST /embeddings	Generate	Gap-filling content
5	Sora Video Generation → Analysis Loop	POST /images/generations (Sora)	Mave Agent (Chat)	Iterated video with quality scores

Rate Limits & Production Notes

OpenAI Endpoint	Rate Limit	Strategy
`/audio/transcriptions` (Whisper)	50 RPM	Queue audio files; 2s delay between calls
Responses (GPT-4.1 vision)	500–10,000 RPM (tier-dependent)	Batch frames; use `detail: "low"` to reduce tokens
`/audio/speech` (TTS)	50 RPM	Sequential with 1s delay
`/embeddings`	3,000 RPM / 1M TPM	Batch up to 2,048 inputs per request
`/images/generations` (Sora)	Tier-dependent	Single request with retry; expect 30-120s

Use OpenAI’s usage dashboard to monitor token and credit consumption in real time. Set monthly spend limits under Organization → Billing → Usage limits to prevent unexpected charges.

Rate limits vary by organization tier. New accounts start with lower limits that increase with usage history. For 429 errors, use exponential backoff: wait 1s, 2s, 4s, 8s between retries, capped at 60s. Whisper files over 25 MB must be split. TTS inputs over 4,096 characters must be chunked at sentence boundaries. Monitor Mavera credits at Dashboard.

​API Reference Card

​Prerequisites

​Jobs

​Rate Limits & Production Notes

All Integrations

OpenAI API Docs

Mave Agent

Generate

Brand Voice

Focus Groups

API Reference Card

Prerequisites

Jobs

Rate Limits & Production Notes