# MusicGPT API > API from MusicGPT in Singapore for generating songs, vocals and sound effects from prompts, with remixing, extending, covers, voice conversion, text to speech, stem extraction and audio utilities. Jobs are asynchronous, with results by webhook or polling. - Canonical: https://www.anchorterminal.com/tools/musicgpt - Markdown: https://www.anchorterminal.com/tools/musicgpt.md (~6,750 tokens) - Slim: https://www.anchorterminal.com/tools/musicgpt.min.md (~1,630 tokens, same facts, less prose, for token-sensitive contexts) - JSON: https://www.anchorterminal.com/tools/musicgpt.json (this page as data, same URL with Accept: application/json) - Site index for agents: https://www.anchorterminal.com/llms.txt (full text: https://www.anchorterminal.com/llms-full.txt) - API: https://www.anchorterminal.com/api/v1/index.json - Updated: 2026-10-10 ## Overview **Grade E · 45.9/100 · rank #859 of 950 · #6 in Music generation · not agent-ready · confidence medium** ## Assessment A public OpenAPI 3.1 file with 27 operations, llms.txt and per-request prices cover song generation, editing, stems and sound effects under one key. No status page, SLA, security.txt or idempotency key was found, and the privacy policy says prompts and uploads may be used to train models. ## Facts | Field | Value | | --- | --- | | Vendor | MusicGPT Ltd (https://musicgpt.com) | | Kind | Model API | | Category | Music generation (https://www.anchorterminal.com/categories/music-generation) | | Transport | HTTP | | Endpoint | `https://api.musicgpt.com/api/public` | | Auth | API key · Self-serve API key. A person signs in at musicgpt.com (email code or Google, Facebook, Apple or Discord login), creates a named key in the API dashboard and sends it as the raw `Authorization` header value. No scopes, expiry or rotation are documented. | | Pricing | Pay per use ($0.10 / call) · Per request from prepaid credits. Pay as you go is $0.10 a song on `v6`. API Plus ($19 a month) and API Pro ($69 a month) lower the rates and include no credits, and `v7` models need Pro. Music requests return two songs and bill each. The pricing page says up to 10 free song generations, while the account page asks for payment information (https://docs.musicgpt.com/api-documentation/index/pricing). | | x402 | No · No x402, MPP or L402 in the docs, the OpenAPI file or the pricing page (checked 2026-10-09). | | Licence | Proprietary service under MusicGPT's Terms of Service. The npm client is ISC | | Packages | npm: `musicgpt`; pypi: `musicgpt` | | Docs | https://docs.musicgpt.com | | llms.txt | https://docs.musicgpt.com/llms.txt | | npm downloads / week | 310 | | Models | `v6`, `v6-pro`, `v7` and `v7-pro` on `POST /v2/MusicAI`, chosen by `model` (default `v6`). The pro models stream. Music AI V1 is deprecated since June 2026 | | Track length | No maximum song length published. Sound Generator takes `audio_length` from 5 to 120 seconds. Input audio up to 4 minutes for Remix, Extend, Inpaint and Sing Over Instrumental, 8 minutes elsewhere | | Output | MP3, and WAV for Music AI, Remix, Extend, Inpaint and Sing Over Instrumental. Two versions a music request. MP3 files kept 90 days and WAV files 7 days | | Stems | `POST /v2/Extraction` with 23 stem names, from vocals and instrumental to kick drum and hi-hat, priced per request plus per minute | | Vocals | Yes, with lyrics up to 5,000 characters, a `voice_id` from the voice lists, instrumental-only and vocal-only modes | | Licence of output | The terms give users on paid plans full commercial rights. MusicGPT takes a perpetual, non-exclusive licence to use generated music for service improvement, marketing and training | | Training data | Not described for the models. The privacy policy says prompts and uploads may be used for training after anonymisation | | Free tier | Up to 10 free song generations per the pricing page. Sound Generator is free during October 2026, up to 500 a day and 21 an hour | | Rate limits | Free 1, Plus 3, Pro 10 parallel requests per feature group. `GET /v1/byId` 20, 200 or 500 calls a minute | | Job model | Create returns `task_id`, `eta` and `credit_estimate`. Results by `webhook_url` or `GET /v1/byId`. No webhook signature documented | | Errors | `success`, `error_code` and `message`, 11 codes from 400 to 503, with 402 `INSUFFICIENT_CREDITS` and 429 `TOO_MANY_REQUESTS` | | SDKs | npm `musicgpt` 1.0.1 (16 June 2025, ISC, TypeScript types) and PyPI `musicgpt`. A Zapier app is linked | | Capabilities | audio.music, audio.vocals, audio.stems, audio.sfx | | Tags | hosted, closed-source, commercial-licence, async-jobs, webhooks, openapi, llms-txt, typescript, python, free-tier | | JSON | https://www.anchorterminal.com/api/v1/tools/musicgpt.json | ## Score breakdown (methodology v0.4, October 2026 research run) Assessed 2026-10-09 from public evidence against the published checklist (https://www.anchorterminal.com/benchmark/#checklist). Confidence: medium. Performance and Task success pending (no score, not in the total); the total is Σ(score × weight) ÷ 80 over the 7 assessed categories. "This run" is each category's share of the 100 points. | Category | Weight | This run | Score (0–100) | Points | | --- | --- | --- | --- | --- | | Reliability | 16% | 20 | 37 | 7.4 | | Performance | 10% | pending | pending | n/a | | Schema & documentation | 13% | 16.2 | 78 | 12.7 | | Agent ergonomics | 13% | 16.2 | 54 | 8.8 | | Security & auth | 14% | 17.5 | 35 | 6.1 | | Payments & pricing | 10% | 12.5 | 30 | 3.8 | | Task success | 10% | pending | pending | n/a | | Maintenance & community | 7% | 8.8 | 33 | 2.9 | | Transparency & trust (editorial 32, provenance 65) | 7% | 8.8 | 49 | 4.3 | | Negative events | up to −15 | up to −15 | none recorded | 0 | | **Total** | | | | **45.9 → E** | ### Why each score - Reliability 37: Graded as a hosted API. No status page is linked from the docs, the home page or the API page (0). No readable incident history (5). Limits are published per plan, 1, 3 or 10 parallel requests per feature group and 20, 200 or 500 `GET /v1/byId` calls a minute (15). The error page names 429 `TOO_MANY_REQUESTS` and asks for exponential backoff, with no Retry-After header and no idempotency key documented (7). No SLA found. The API page claims 99 per cent server uptime with no terms behind it (0). `POST /v2/MusicAI` is not labelled beta (10). - Performance: Pending. Latency is measured per call by our probes, which haven't run yet, so this run doesn't score it. Its weight is shared across the assessed categories until the first probe window closes. - Schema & documentation 78: A public OpenAPI 3.1 file with 27 operations (25). llms.txt and a Markdown copy of every page (10). Each operation has a one-line purpose and the guides add notes such as `voice_id` forcing the `v6` model, with little on when not to use an endpoint (11). Schemas are inline with no shared components, 12 of 23 request bodies name required fields, and `gender` on `/v2/MusicAI` is a free string although the guide lists three values. Stems travel as a JSON string inside a form field (9). Every operation has an example, and the error page lists 11 codes, while the file documents only 402, 422 and 500 (11). `/v1` and `/v2` paths and a changelog dated by month (12). - Agent ergonomics 54: Creates return a small JSON task with an `eta` and a `credit_estimate`, but music requests always make two versions and Music AI V2 sends up to seven webhooks a request (15). `limit` and `page` on the two voice lists, polling by `task_id` or `conversion_id`, and no call that lists past conversions (10). One error body with `error_code` and `message` and 11 documented codes (16). No idempotency key is documented and 409 `CONFLICT` is described only in general terms (3). `/v2/MusicAI` has no required field and defaults to `v6`. SDKs are published for Node and Python, and the npm package dates from 16 June 2025 (10). - Security & auth 35: Named API keys created in the API dashboard and sent as the raw `Authorization` header value. No scopes, expiry or rotation are documented (20). No read-only or restricted key (0). Output is generated audio, lyrics and transcripts of the caller's own files, so no third-party content (10). Each conversion can be fetched by ID with its `conversion_cost`, and no audit log was found (5). `/.well-known/security.txt` answers 404, and no disclosure policy, bug bounty or certification was found on the pages read. No webhook signature is documented (0). - Payments & pricing 30: No x402, MPP or L402 in the docs, the OpenAPI file or the pricing page (0). Per-request prices for every feature on three plans are public (20). The pricing page says up to 10 free song generations and the rate-limit page has a Free plan column, while the account page says to enter payment information, so the card question is open and the line takes half (10). A person signs up in a browser by email code or social login and creates the key in a dashboard (0). - Task success: Pending. Task success needs the category task suites run through each tool, which haven't run yet, so this run doesn't score it. Its weight is shared across the assessed categories until then. A data provider's data-quality score is published on its listing now and becomes half of this category when it's scored. - Maintenance & community 33: The newest changelog entry is labelled September 2026 with no day, so recency is scored within 90 days (20). Two entries fall in the last 90 days, August and September 2026 (0). A public changelog and an API support address with a stated 24 hour response, with no public issue tracker (8). Official SDKs exist on npm and PyPI, but npm `musicgpt` 1.0.1 dates from 16 June 2025 and names no repository (5). No public repository or CI to judge (0). - Transparency & trust 49: Closed service under one 14-section Terms of Service dated 1 January 2025 that covers the whole platform, with no API-specific terms. Paid plans carry commercial rights to output (12). The privacy policy is ten short sections. It says prompts and uploads may be used to train models after anonymisation and gives no retention periods. The changelog says MP3 files are kept 90 days and WAV files 7 days. No DPA found (10). The changelog gives dated notices, such as WAV files leaving webhooks on 9 February 2026 and new prices from 1 August, with no written policy and no removal date for Music AI V1 (10). No sub-processor list or data locations found (0). Fix list for a coding agent, everything this grade says the listing lacks, the biggest gain first (21 items): https://www.anchorterminal.com/fixes/musicgpt.md (JSON https://www.anchorterminal.com/fixes/musicgpt.json) ### What we couldn't check - unchecked: the PyPI project page for `musicgpt` answered with a client challenge, so the Python SDK's version and release date are unknown. - unchecked: the API dashboard is behind a sign-in, so key revocation, usage logs and whether the free allowance needs a card were not seen. - unchecked: the 22 feature guides other than Music AI V2 and the prompt guideline were not read. Their parameters were taken from the OpenAPI file. - Whether a status page exists. None is linked from the docs, the home page or the API page. - The changelog of January 2026 mentions song length control in Music AI. No length parameter appears in the `/v2/MusicAI` schema or guide, and no maximum song length is published. - The terms name MusicGPT Ltd. Singapore with no registration number or address. The npm maintainer account and a sample audio bucket carry the name Lalals, and the relation was not established. - The terms bar building scrapers or similar programs to gain access to MusicGPT's technology. Recorded as a fact with no deduction. It matters before any probe is run. - The API page carries a paragraph of placeholder Latin text and claims 99 per cent uptime and latency under 180 ms with no source. - Process notes. robots.txt was read first on musicgpt.com and docs.musicgpt.com (both 200, both allow the pages read). registry.npmjs.org answered the robots.txt path with a package document and pypi.org allows `/project/`. The robots.txt requests to rdap.verisign.com (400) and api.npmjs.org (404) were sent in one command, which the brief asks to keep apart. ### Sources - docs index (llms.txt): (seen 2026-10-09) - OpenAPI 3.1 file, read as the description file and not the rendered reference pages: (seen 2026-10-09) - pricing: (seen 2026-10-09) - rate limits: (seen 2026-10-09) - error handling: (seen 2026-10-09) - changelog: (seen 2026-10-09) - authentication and base URL: (seen 2026-10-09) - API key creation: (seen 2026-10-09) - account creation: (seen 2026-10-09) - Music AI V2 guide: (seen 2026-10-09) - webhooks: (seen 2026-10-09) - SDK and tools: (seen 2026-10-09) - FAQs: (seen 2026-10-09) - API product page: (seen 2026-10-09) - Terms of Service: (seen 2026-10-09) - Privacy Policy: (seen 2026-10-09) - about page: (seen 2026-10-09) - security.txt (404): (seen 2026-10-09) - npm package metadata: (seen 2026-10-09) - domain registration (RDAP): (seen 2026-10-09) ## Who's behind it (provenance 65/100, checked 2026-10-09) | Check | Finding | Points | | --- | --- | --- | | Legal entity named | MusicGPT Ltd | 20/20 | | Domain age | musicgpt.com, registered 2021-11-04 (4 years) | 7/15 | | Endpoint on the vendor's domain | api.musicgpt.com | 15/15 | | Terms of service | read, states 5 of the 7 things a reader expects, and has 1 clause that costs points | 6.3/10 | | Privacy policy | read, states 6 of the 8 things a reader expects, and has 1 clause that costs points | 6.5/10 | | Status page | not found | 0/10 | | Changelog | published | 10/10 | | security.txt | not found | 0/10 | The Terms of Service (effective 1 January 2025) name MusicGPT Ltd. Singapore and are governed by Singapore law. No registration number or address is given. One Terms of Service and one Privacy Policy cover the whole platform. No API-specific terms or DPA were found. The docs give the base URL https://api.musicgpt.com/api/public. No request was sent to it. musicgpt.com/.well-known/security.txt answers 404. RDAP for musicgpt.com gives a registration date of 2021-11-04 and GoDaddy.com, LLC as registrar. The npm maintainer account uses a lalals.com address and a docs sample is served from a bucket named lalals. The relation to MusicGPT Ltd was not established. ### Terms and privacy, as read A reading by a fixed set of rules, each answered with the vendor's own sentence. Not legal advice. **Terms of service** (https://musicgpt.com/terms-of-service), read 2026-10-09, dated 2025-01-01, states 5 of the 7 things a reader expects. - To know. Restricts automated access (costs points). "Build scrapers or similar programs in an attempts to gain access to MusicGPT's technology." - To know. Says access can be ended without notice or for any reason. "We reserve the right to terminate or suspend your access to the Service at our discretion, particularly for violations of these Terms or misuse of the platform." - Gives the date it was last updated. Last updated 2025-01-01. - Names the governing law or courts. The law of Singapore. - Not found in the text. Says how changes to the terms are announced. - Not found in the text. Refers to a service level or uptime commitment. - Also in the text (2026-10-08). Users grant MusicGPT a perpetual, royalty-free licence to use, showcase and distribute generated music for improving the service, marketing and training. "By using the Service, you grant MusicGPT a royalty-free, non-exclusive, and perpetual license to use, showcase, and distribute generated music for improving the Service, marketing, and training purposes." - Also in the text (2026-10-08). The Ultra Plan is stated as not intended for automated systems, large-scale data processing or reselling. "The Ultra Plan is designed for normal professional and creative use and is not intended for automated systems, large-scale data processing, reselling, or other atypical workloads that could cause disproportionate strain on our systems." - Also in the text (2026-10-08). Full commercial rights to generated content are stated for users on paid plans. "Users on paid plans have full commercial rights to their generated content." **Privacy policy** (https://musicgpt.com/privacy-policy), read 2026-10-09, dated 2025-01-01, states 6 of the 8 things a reader expects. - To know. Says it may use customer content to train or improve models, and no opt-out was found (costs points). "Your prompts and uploads may be used to train our AI models." - Gives the date it was last updated. Last updated 2025-01-01. - Not found in the text. Says whether personal data is sold or shared for advertising. - Gives a privacy contact. support@musicgpt.com. - Not found in the text. Says where data is transferred or stored. ## Live (updated 2026-10-10 04:40 UTC) - Right now: up, HTTP 404, 356 ms, checked 2026-10-10 04:40 UTC (get on `https://api.musicgpt.com/api/public`) - Uptime 24h 100.0% (133 probes) · 30 days 100.0% (133 probes) · p50 347 ms · p95 457 ms - npm `musicgpt` 1.0.1 - pypi `musicgpt` 0.0.0, released 2025-05-16 - Watching changelog - Watching pricing - Watching privacy - Watching terms - Always current: https://www.anchorterminal.com/api/v1/live/musicgpt.json ## Probe metrics Not measured yet. Our benchmark probes haven't run, so there's no availability, latency or error rate from a run and Performance is pending. Live uptime, where we poll the endpoint, is under Live and doesn't change the score. ## Prices | Item | Price | Unit | Note | | --- | --- | --- | --- | | AI Music Generator, v6, pay as you go | $0.10 | per call | per song, two songs a request by default | | AI Music Generator, v6, Pro plan | $0.035 | per call | per song, on top of $69 a month | | AI Music Generator, v7-pro, Pro plan | $0.06 | per call | per song, Pro plan only | | Sound Generator, pay as you go | $0.028 | per call | per generation, free during October 2026 | | API Plus plan | $19 | per month (plan) | lowers per-call rates, no credits included | | API Pro plan | $69 | per month (plan) | lowest per-call rates, no credits included | Across all listings: https://www.anchorterminal.com/prices/index.md ## Strengths - OpenAPI 3.1 file with 27 operations, plus llms.txt and a Markdown copy of every docs page - Per-request prices for every feature are published without a login, from $0.035 to $0.10 a song on the `v6` model - Rate limits are published per plan, 1, 3 or 10 parallel requests and 20, 200 or 500 status calls a minute - One error body (`success`, `error_code`, `message`) with 11 documented codes - Results arrive by webhook or by polling `GET /v1/byId`, and each response carries a cost figure ## Weaknesses - No status page, SLA or incident history was found on the pages read - No idempotency key is documented, and music requests bill two songs each by default - The privacy policy says prompts and uploads may be used to train models, with no opt-out stated - No security.txt (404), disclosure policy, certification or webhook signature was found - The npm SDK was last published on 16 June 2025, before the `model` parameter was added in June 2026 - The terms forbid building scrapers or similar programs to gain access to MusicGPT's technology. This matters before any probe is run ## Before you call it (notes for agents) 1. Send the key as the raw `Authorization` header value. The docs show no `Bearer` prefix 2. Budget two songs for each music request. `POST /v2/MusicAI` returns two versions and the price is per song 3. Save `task_id` and poll `GET /v1/byId` with `conversionType`, or set `webhook_url`. Music AI V2 sends up to seven webhooks a request 4. Pass `model` explicitly. The default is `v6`, and `voice_id` or `vocal_only` forces `v6` whatever is passed 5. Do not retry a create blindly after a timeout. No idempotency key is documented, so a retry may bill again ## Connect Install: ```bash npm install musicgpt ``` First request: ```bash curl -X POST https://api.musicgpt.com/api/public/v2/MusicAI -H "Authorization: $MUSICGPT_API_KEY" \ -H "Content-Type: application/json" \ -d '{"prompt":"a song about greenery, nature and forest","model":"v7-pro"}' ``` ## Similar tools Ranked by shared capabilities, then score. Same-category tools with no shared capability key are listed last. | Tool | Grade | Score | Rank | Shared capabilities | x402 | Markdown | | --- | --- | --- | --- | --- | --- | --- | | fal music models | B | 68.3 | 222 | audio.music, audio.stems, audio.sfx, audio.vocals | no | https://www.anchorterminal.com/tools/fal-music.md | | Soundverse API | E | 41.9 | 898 | audio.music, audio.stems, audio.sfx, audio.vocals | no | https://www.anchorterminal.com/tools/soundverse.md | | ElevenLabs Music API | B | 66.9 | 268 | audio.music, audio.vocals, audio.stems | no | https://www.anchorterminal.com/tools/elevenlabs-music.md | | Loudly Music API | E | 38.1 | 923 | audio.music, audio.stems, audio.vocals | no | https://www.anchorterminal.com/tools/loudly.md | | Stable Audio API | D | 50 | 782 | audio.music, audio.sfx | no | https://www.anchorterminal.com/tools/stable-audio.md | | Google Lyria | D | 48.5 | 811 | audio.music, audio.vocals | no | https://www.anchorterminal.com/tools/google-lyria.md | ## Panel reviews (0) Reviewed by the Anchor panel (https://www.anchorterminal.com/reviewers/index.md): . Desk reviews, written from public documentation, pricing, terms, source and status history on 1 October 2026. No calls made. For a desk review, the outcome says whether the reviewer's questions could be answered from public material: success, partial or failure. How reviews work: https://www.anchorterminal.com/reviews/how-it-works.md ## Notable - Four music models (`v6`, `v6-pro`, `v7`, `v7-pro`) are chosen per request with `model` on `POST /v2/MusicAI`, and the two pro models stream (source: ) - Plans lower the price of each call and include no credits. Pay as you go draws down credits added up front (source: ) - The OpenAPI 3.1 file lists 27 operations, among them remix, extend, inpaint, cover, voice changer, text to speech, stem extraction, transcription and audio to MIDI (source: ) - The changelog of August 2026 marks `v2/MusicAI/stream` and `v2/MusicAI/batch` deprecated, and June 2026 deprecates Music AI V1 (source: ) - The privacy policy says prompts and uploads may be used to train AI models after anonymisation (source: ) - The terms forbid building scrapers or similar programs to gain access to MusicGPT's technology (source: ) - #6 of 11 in Best music generation for AI agents: https://www.anchorterminal.com/best/music-generation/index.md - All 55 music comparisons: https://www.anchorterminal.com/compare/music-generation/index.md ## Compare - [Beatoven.ai API vs MusicGPT API](https://www.anchorterminal.com/compare/beatoven-vs-musicgpt.md): F 22 vs E 45.9 - [ElevenLabs Music API vs MusicGPT API](https://www.anchorterminal.com/compare/elevenlabs-music-vs-musicgpt.md): B 66.9 vs E 45.9 - [fal music models vs MusicGPT API](https://www.anchorterminal.com/compare/fal-music-vs-musicgpt.md): B 68.3 vs E 45.9 - [Google Lyria vs MusicGPT API](https://www.anchorterminal.com/compare/google-lyria-vs-musicgpt.md): D 48.5 vs E 45.9 - [Loudly Music API vs MusicGPT API](https://www.anchorterminal.com/compare/loudly-vs-musicgpt.md): E 38.1 vs E 45.9 - [Mubert API vs MusicGPT API](https://www.anchorterminal.com/compare/mubert-vs-musicgpt.md): E 41 vs E 45.9 - [MusicGPT API vs MusicGen on Replicate](https://www.anchorterminal.com/compare/musicgpt-vs-replicate-musicgen.md): E 45.9 vs C 54.3 - [MusicGPT API vs SOUNDRAW API](https://www.anchorterminal.com/compare/musicgpt-vs-soundraw.md): E 45.9 vs F 12 - [MusicGPT API vs Soundverse API](https://www.anchorterminal.com/compare/musicgpt-vs-soundverse.md): E 45.9 vs E 41.9 - [MusicGPT API vs Stable Audio API](https://www.anchorterminal.com/compare/musicgpt-vs-stable-audio.md): E 45.9 vs D 50 ## Verify this listing For the vendor. The badge or a plain link to this page verifies the listing, from a page on musicgpt.com or one of its subdomains. It shows the listing is the vendor's and that the vendor knows it's here, and it never changes a grade, rank or review. The vendor sends the page's address to `POST https://www.anchorterminal.com/api/v1/verify` as `{"slug": "musicgpt", "url": "…"}`, or calls the `verify_listing` tool at https://www.anchorterminal.com/mcp. We fetch the page once, then again every week; two failed checks in a row and the verification lapses, and a later pass restores it. What we check: https://www.anchorterminal.com/builders/index.md#verify HTML badge: ```html MusicGPT API on Anchor Terminal ``` Markdown badge, for a README: ```markdown [![MusicGPT API on Anchor Terminal](https://www.anchorterminal.com/badges/musicgpt.svg)](https://www.anchorterminal.com/tools/musicgpt) ``` Plain link: ```html MusicGPT API on Anchor Terminal ``` ## Share this listing For the vendor. Sharing assets for social media, two PNGs of 1200 × 630 that say MusicGPT API is listed on Anchor Terminal, with the vendor's logo and this page's address and no grade or score. - Dark: https://www.anchorterminal.com/assets/share/musicgpt-dark.png - Light: https://www.anchorterminal.com/assets/share/musicgpt-light.png