How do I get an API key?

Sign up for any plan (the free tier includes API access), then create a key from the dashboard under Settings → API Keys. Keys are prefixed ap_ and bound to your account.

What's the difference between /developers and the API docs?

This page is the marketing-side entry point — the elevator pitch and 'what fits where'. Full reference, every endpoint, every parameter, lives at docs.audiopod.ai.

Is there a free tier for development?

Yes. The free plan includes API access with a per-day credit allowance. Upgrade to paid plans for higher throughput and premium features.

Which AI agent frameworks does the MCP server support?

Anything that speaks MCP — Claude Desktop, Cursor, Continue, Cline, Goose. Drop in mcp.audiopod.ai and authenticate with your API key.

How does pricing work for API calls?

The API runs on a separate, prepaid API wallet billed in USD — distinct from the web app's monthly credits. You add funds once and pay per minute of audio processed; failed jobs are refunded and your balance never expires. Your first API key includes $1 of free credit. See the full per-minute rate table and an interactive cost calculator at audiopod.ai/developers/pricing.

Can my AI agent discover AudioPod automatically?

Yes. AudioPod publishes /.well-known/{mcp/server-card.json, oauth-protected-resource, agent-card.json, agent-skills/index.json, api-catalog} plus /llms.txt, /llms-full.txt, and /skill.md so agents can self-onboard.

For developers

AudioPod: the all-in-one audio API.

Six capabilities, one key — speech, transcription, music, stems, voices and more. No card to start ($1 free), and you're never charged for a job that fails. REST API, Python and Node SDKs, and an MCP server for AI agents.

Read the docs

What you get

One stack, every audio capability.

Same primitives. Same auth. Same SDK shape across languages.

REST + MCP

Same capabilities exposed twice — one for HTTP clients, one for AI agents.

Python + Node SDKs

Identical surface across both languages. Same shapes, same async ergonomics.

OAuth + API keys

Per-scope permissions, RFC 9728 discovery, ap_* key format with rate limits.

Streaming + async

Long jobs return job IDs; short jobs stream. SDKs hide the difference.

Try it live — no signup

Hear AudioPod in your browser before you write a line of code.

Type a sentence, hit generate, get audio. The same call works from curl, Python, Node, an MCP-aware editor, or a Claude agent — tabs show you exactly how.

What should AudioPod read aloud?

0 / 200

This call goes to AudioPod with a short-lived demo key. Your text is processed once and not stored.

5-min Quickstart API Reference SDK Reference Recipes API Pricing

Capabilities

Four primitives. Compose anything.

Each capability is reachable from the API, the SDKs, the CLI, and via MCP. Same auth, same patterns.

Synthesize speech

Generate high-quality TTS in 200+ languages. Voice cloning, pause & pronunciation markup, streaming and batch.

Generate music

Full songs from a text prompt. Royalty-free for commercial use on paid plans. Stems on demand.

Transcribe + diarize

99.8% accuracy with speaker labels. SRT, VTT, JSON output for any downstream pipeline.

Split stems

A 45-instrument stem catalog from any track. Vocal removal, instrument isolation, adaptive mixing.

For AI agents

Agent-native by default.

If you're building an AI agent that needs audio, AudioPod is callable out-of-the-box. Drop in the MCP server, or use the SDK from inside your agent loop.

# Drop into Claude Desktop / Cursor / Continue / Cline / Goose
mcp.audiopod.ai

# Or fetch the one-shot skill brief
https://audiopod.ai/skill.md

Troubleshooting

Common API errors, and how to fix them.

The first 401 / 402 / 429 you'll see, plus what the response means and the one-line fix.

Code	Cause	Fix	Docs
401 Unauthorized	Missing or invalid API key. Header expected: Authorization: Bearer ap_…	Create a key in the dashboard (Settings → API Keys). Free tier includes API access.	Open
402 Insufficient credits	Your account ran out of credits for the requested capability.	Top up via PAYG credit packs, or upgrade your plan. Or wait — free tier resets daily.	Open
402 PREMIUM_TIER_REQUIRED	You requested a premium capability variant that your current plan doesn't include.	Either drop the premium-only parameter (the call succeeds at the standard tier), or upgrade your plan.	Open
402 CUSTOM_VOICE_LIMIT_REACHED	You've used every custom voice slot your plan allows. The response detail carries used, limit, and current_plan.	Delete a custom voice you no longer need, or upgrade your plan for more slots (Creator and above are unlimited).	Open
404 voice_not_found	The voice slug or catalog ID you passed isn't in the public voice catalog (or isn't visible to your account).	Browse https://audiopod.ai/voices for the live catalog, or call GET /api/v1/voice/public-voice-profiles to list slugs your key can use.	Open
429 Too many requests	You exceeded the per-minute request quota for your plan tier.	Add exponential backoff with the Retry-After header. SDKs handle this automatically. Long-term: upgrade to a higher-throughput plan.	Open
5xx	Internal error. The request ID lives in the X-Request-Id response header — quote it when reporting.	Retry with backoff. If it persists, check https://status.audiopod.ai or email support@audiopod.ai with the request ID.	Open

Same table is published as JSON at /developers/errors.json for AI agents to cite.

Common questions

Start with the free tier.

No credit card required. Upgrade when throughput matters.

Open the quickstart Create an API key