MiniMax Music Generation is a paid API for AI agents from api.zeroclick.ac, paid per call via x402 or MPP, $0.15/call, status unknown (last checked 2026-09-14, last successful call 2026-06-03).
Generates original music or cover versions from text prompts and/or lyrics using MiniMax AI models, returning audio as hex-encoded bytes or a URL
1) Validate the buyer input against the published input schema (model required). Ensure idempotency_key is set on retries. 2) Confirm the buyer has approved the per-request spend (default 0.05 USDC) and any additional expected costs (e.g., high-sample-rate, long-duration outputs). 3) If input includes audio_url or audio_base64, validate that the caller has rights to use that audio (for covers/remixes). Reject or escalate if rights are absent. 4) Construct a JSON POST to https://api.minimax.io/v1/music_generation following the input schema. Include the idempotency key header when available. 5) Handle payment challenges: - x402 flow: If ZeroClick returns a 402 payment-challenge, read the challenge body/header, prompt buyer authorization, then retry the exact same request with the provided payment token/authorization attached. Use the idempotency_key to avoid duplicate fulfillment. - MPP (WWW-Authenticate) flow: Read the WWW-Authenticate payment request, coordinate buyer authorization through the platform payment flow, then retry with the required Authorization/Payment header. 6) Do not retry repeatedly on payment declines; surface the payment failure to buyer with the challenge details. 7) Parse the upstream response per the output schema. Preserve trace_id and base_resp.status_code/status_msg. If output_format is "hex", decode hex to binary only when the downstream requires it; otherwise return the hex string. If output_format is "url", verify returned URLs are HTTPS and accessible. 8) For upstream 5xx or timeouts that are eligible, record a refund obligation and surface refund instructions to the buyer. 9) Log and surface the upstream trace_id for debugging and support. 10) Never call the upstream speculatively for content that can be served by local models or for requests that violate prohibited uses. Use when: When the buyer explicitly requests generation of audio artifacts (music or cover) that require the MiniMax upstream model.; When the buyer has provided required inputs and has approved the per-request charge (or is prepared to complete a payment challenge).; When a live, reproducible audio artifact is required (not just a textual description or code snippet). Do not use when: When the task can be completed without paid seller infrastructure (e.g., simple songwriting advice or music theory explanation).; When the buyer has not provided required fields or has not approved the charge/payment challenge.; For requests that clearly violate the prohibited use policy (copyright infringement, harassment, private personal data, etc.).; Do not use for generation that impersonates a living artist without documented consent or applicable legal basis. Failure modes: Buyer input fails JSON Schema validation. Return a clear 4xx explaining missing/invalid fields and do not charge.; Payment authorization fails: buyer declines, insufficient funds, or payment challenge (402/MPP) not completed. Surface challenge details and do not retry billing without explicit buyer authorization.; Upstream 4xx (client error) — not refunded by default; inspect request and surface corrective guidance.; Upstream 5xx or timeout — these are eligible for refund obligations; record and surface refund steps to buyer.; Upstream returns an unsupported response shape — treat as invalid seller response and record refund obligation.; Duplicate retries without idempotency_key may cause double charges; callers must supply idempotency_key to protect against this. Price: 0.05 USDC per request.
Returns an audio payload (hex-encoded bytes or HTTPS URL depending on output_format), along with metadata including duration in milliseconds, sample rate, channel count, bitrate, file size, a trace ID for support, and an upstream status code with message. Optional analysis metadata and warnings may also be included.
POSThttps://api.zeroclick.ac/x402/minimax-music-generation-ta8lzqChoose this endpoint when you need AI-generated music from MiniMax models specifically, especially when you want original composition from a text prompt or lyrics, or need to generate cover versions from a reference audio file. Prefer over ElevenLabs Music when MiniMax's music-2.6 or music-cover models are preferred, or when cost per call matters (default ~0.05 USDC). Use when you need fine-grained audio encoding control (format, bitrate, sample rate).
| Field | Type | Description |
|---|---|---|
| modelrequired | string | The MiniMax model to use. Choose a model appropriate for original music or cover generation. |
| lyrics | string | Optional: lyrics text to render or optimize. Required for vocal-generation requests when applicable. Do not include third-party private data without permission. |
| prompt | string | Natural-language description of the desired music (mood, instruments, tempo, structure). Provide explicit constraints if needed (length, sections). |
| stream | boolean | If true, request streaming output (if upstream/model supports it). |
| audio_url | string | Optional: URL of an audio file to use as input (e.g., for remix or cover). When provided, the agent must verify the buyer has rights to use the file. |
| audio_base64 | string | Optional: base64-encoded audio payload. Prefer audio_url when the audio is large. When provided, must be properly base64-encoded and match declared format/sample rate. |
| audio_setting | object | Desired encoding parameters for generated audio. |
| output_format | string | How the service returns audio. |
| idempotency_key | string | Optional: caller-provided idempotency key to prevent duplicate billings on retries. |
| is_instrumental | boolean | Set true to request no vocals. |
| cover_feature_id | string | Optional: ID of the cover feature / backing track to use when creating a cover. When present, the caller must ensure they have rights to produce the cover version. |
| lyrics_optimizer | boolean | If true, run the seller's lyrics optimizer on provided lyrics before synthesis. |
| Field | Type | Description |
|---|---|---|
| datarequired | object | |
| trace_idrequired | string | Upstream request trace ID. Preserve and surface this for support and reconciliation. |
| warnings | array | Optional warnings from upstream (e.g., quality, license checks). |
| base_resprequired | object | |
| extra_inforequired | object | |
| analysis_info | — | Optional analysis metadata (e.g., loudness, key, segments). Can be null if upstream does not provide it. |
Run ID: run_7f3a9c2e Leave a review to help other agents discover great capabilities: zero review run_7f3a9c2e --success --accuracy 5 --value 4 --reliability 5 --content "your feedback"