Murf AI

Published September 3, 2026 · This page indexes everything this site has verified about Murf and links to where each finding is set out in full. It is not a review — the review is. No listening test has been run, so nothing here says how Murf sounds.

On this page

What it costs

Murf Studio sells hours of Voice Generation Time across four paid tiers — Creator Lite $29/month for 2 hours, Creator Plus $49 for 4, Business Lite $99 for 8, Business Plus $199 for 20 — which the site normalizes to $0.13–$0.24 per minute of generated audio, while annual figures conflict between Murf’s live pricing endpoint ($276/$468/$948/$1,908, a 20.1–20.7% saving) and the pricing page’s own JavaScript bundle ($228/$396/$792/$1,596, 32.7–34.5%), all read August 11, 2026 (the review). The API bills characters instead — Falcon 2 at $0.01 per 1,000 characters and Gen 2 at $0.03, which the pricing benchmark lists as $10.00 and $30.00 per million characters on rates read September 1, 2026 (pricing comparison) — with a minimum purchase of either $2 or $10 depending on which help page you read, plus a 100,000-character free API trial; Murf Dub is a third meter at $1 per credit, 2 credits per minute per language ($2.00/minute) or 62 credits for QA-assured dubbing ($62.00/minute), read August 11, 2026. Murf publishes no characters-to-minutes conversion for any product, so the site prints no Studio cost per 1,000 characters and no API cost per minute; and because both murf.ai/pricing and murf.ai/api/pricing serve an empty container — the served HTML of the first containing zero dollar-sign characters — every Murf figure the site publishes comes from Murf’s own JSON pricing endpoints (the API one returning "pricingTestGroup":"B") or from server-rendered help-centre prose.

Every rate above is set against every other provider’s, normalized to one unit, in our pricing comparison.

Where Murf’s own material disagrees with itself

These are contradictions inside the vendor’s published documentation, not disagreements between us and the vendor. We report both readings rather than choosing one, and each links to the page where we set it out in full. We have published 36 in total; the 10 with the most bearing on a decision are below, and the rest sit in the pages linked underneath.

• Five different strings circulate for what appears to be one streaming model: "Falcon 2" and falcon-2 in the API documentation, "Falcon2" on the WebSockets reference, "Falcon" and "Murf Falcon" across all marketing pages, and FALCON in the January 2026 changelog and the Agora sample configuration. On August 11, 2026 the string "Falcon 2" appeared zero times on murf.ai, murf.ai/api and murf.ai/falcon, while "Murf Falcon" appeared 6, 9 and 18 times, and Murf publishes no statement reconciling the two names. Where we published this.
• Murf publishes two official answers for its own API rate and minimum purchase. One help page states a flat rate — "you can purchase 1000 characters at $0.03 (minimum purchase amount is $2)" — and never mentions a model split; another states a $10 minimum and splits by model, "Gen 2 model: $0.03/1000 characters. Falcon model: $0.01/1000 characters." A third Murf page independently supports the Falcon figure. Both were live in the same help-centre section on September 1, 2026. Where we published this.
• The same model is priced in two incompatible units. Murf’s documentation and help centre price Falcon 2 at $0.01 per 1,000 characters; its marketing pages price the same model at "1 cent per minute". No characters-per-minute constant is published anywhere, so the two figures cannot be reconciled. Where we published this.
• Murf’s own pricing payload sets charactersPerUnit to 10,000 for pay-as-you-go, while the pricing page’s code labels the same rates "per 1,000 characters". Read against 10,000 characters the payload reconciles with the help centre; read against the label the page code attaches, the same numbers would render ten times higher. Where we published this.
• Two different annual Studio price sets ship in Murf’s own production assets. The live pricing endpoint serves $276 / $468 / $948 / $1,908; the pricing page’s own JavaScript bundle hardcodes $228 / $396 / $792 / $1,596. Monthly prices are identical in both. The endpoint prices work out to a 20.1–20.7% annual saving and the bundle prices to 32.7–34.5%, while the help centre claims "a flat 33% off" — matching the bundle, not the endpoint. Where we published this.
• Murf publishes two latency figures for the same family. Its API documentation gives Falcon 2 a time-to-first-audio of "~100ms" and "under 100ms"; its marketing pages give "55ms model inference", "130ms end-to-end latency" and "130ms time-to-first-audio consistently across geographies". Only the time-to-first-audio figures are comparable, and they differ. Where we published this.
• Murf defines the same latency metric two different ways. Its latency documentation names the metric "Time to First Audio Byte" and operationalises it against the curl field time_starttransfer; its benchmark page defines time-to-first-audio as the delay between a synthesis request and the first audio frame. A byte and a frame are not the same event, so two Murf pages using one acronym are measuring two different things. Where we published this.
• Three different pronunciation-accuracy figures circulate. The API documentation states 99.37% for Falcon 2; the same 99.37% also appears on the Gen2 model page and in the November 2025 Falcon (Beta) changelog, so one accuracy result is attributed to three different model objects. Murf’s marketing pages print 99.38%, and its Gen 2 marketing page prints a third figure, "over 98.8% word level pronunciation accuracy", from a different test. Where we published this.
• Terms of Service §5.1 grants commercial use with no plan, tier or payment qualifier anywhere in the clause, while Murf’s help centre scopes the same grant to "all paid plans". Both were read on August 11, 2026; the site states the disagreement rather than harmonizing it, and extends neither reading to the free trial. Where we published this.
• The most consequential conflict on Murf’s site, in the site’s own words: Terms of Service §4.2.5 prohibits submitting any recording containing the voice of a speaker who has not explicitly consented in writing, while the FAQ on Murf’s own voice-cloning marketing page answers "Can I clone a celebrity voice?" with "Yes", followed by an ethical caveat rather than a prohibition. Both documents were live the same day; a reader acting on that "Yes" would breach §4.2.5 and separately §11.2.11. Where we published this.

Dated and time-sensitive

Facts with a clock on them — model sunsets, policy revisions, market status. These are the entries most likely to be out of date first, which is why they carry their dates.

• 2026-07-20. Murf’s changelog entry headlined "Upcoming Gen2 Streaming Deprecation" states: "Our legacy Gen2 streaming model is set to be deprecated on 16th August 2026. We recommend migrating to our latest Falcon 2 model promptly. Gen2 non-streaming text-to-speech will remain available through the Synthesize Speech endpoint." The notice appears in eight places in three wordings; the date is identical in all eight. Scope is the streaming path only. Source on this site.
• 2026-08-16. Documented deprecation date for the legacy Gen2 streaming model. Murf uses only the word "deprecated" — never "sunset", "retire", "end-of-life" or "removal" — and publishes nothing on hard failure, error codes, grace period or fallback, nor what the US-East streaming default becomes, US-East being the one region whose Stream Speech default was the model being deprecated. Source on this site.
• 2026-09-01. Re-check after the deprecation date passed: Murf’s model page and streaming guide describe the Gen2 streaming model as deprecated in the present tense, while its OpenAPI specification still lists gen2 as an accepted streaming value. No removal notice and no changelog entry on or after August 16, 2026, so the site does not say the model has been switched off. Advice given: set the model parameter explicitly. Source on this site.
• 2026-09-01. Correction logged on the Murf review: the Gen2 streaming deprecation had been described as forthcoming and its date has now passed; the quick verdict and API section were updated to the current documented position. Murf’s terms-of-service URL was also updated to its current location after the old address began redirecting. All other facts remain as verified August 11, 2026. Source on this site.
• 2026-09-01. Flagship change log, links only: nine Murf AI citations pointed at the old terms-of-service address, which now redirects, and were restated to murf.ai/legal/terms-of-service. No figure, quotation or ranking changed; all facts remain as verified August 11, 2026. Source on this site.
• 2026-06-16. Falcon 2 launch changelog. Murf states "Falcon 2 is now available on: Streaming API, WebSocket API" — it is not available on the non-streaming Synthesize Speech endpoint, whose modelVersion accepts only GEN2. The launch note offers four qualitative bullets with no before/after figures and never names Falcon (Beta) as the thing improved on. Source on this site.
• 2025-11-06. Falcon (Beta) launched, claiming sub-130 ms time-to-first-audio and 99.37% pronunciation accuracy. As of August 11, 2026 it is gone from current documentation with no deprecation notice published; its documentation page 308-redirects to the Falcon 2 page, and no changelog entry announces its retirement. Source on this site.
• 2026-03-05. The request field multiNativeLocale was deprecated, superseded by locale, with no removal date published. Murf documents no behavioural change: the deprecated field and its replacement carry an identical description. Source on this site.

Every page on this site about Murf

• Full review. Murf AI Review 2026: Models, Pricing and Licensing, Verified — the complete documentation-based assessment, with sources and dates against every claim.
• Alternatives. Murf AI alternatives — organized by the reason you would leave, not as a ranked list.
• Head to head. ElevenLabs vs Murf AI, Google Cloud vs Murf AI.
• In our buying guides. Best AI Voice Generators in 2026: 10 Providers Compared, Best AI Dubbing Software 2026: Only Four Vendors Actually Sell It, Best Free AI Voice Generators 2026: What “Free” Actually Buys You, Most Realistic AI Voices: What Every Vendor Charges for the Claim, Best Multilingual AI Voice Generators: Every Count, Recounted, Real-Time AI Voice: Why the Latency Numbers Cannot Be Compared, Best AI Voice Cloning Software 2026: Price, Access and Consent Compared.
• In our comparisons of published data. AI Voice Generator Pricing Comparison: Every Rate Normalized, Best Text-to-Speech APIs: Limits, Transports and Auth Compared.
• In our explainers. Can You Use AI Voices Commercially? What Ten Vendors Actually Say, How AI Voice Cloning Works: What Ten Vendors Require of You, How AI Voice Generators Work, According to the Vendors Themselves.
• By what you are making. AI Voice for Audiobooks: Will Chapter 20 Match Chapter 1?, AI Voice for Business: What Survives a Procurement Review, AI Voice for eLearning: Courses Change, and the Audio Must Follow, AI Voice for Podcasts: Only Three Providers Do Two Speakers, AI Voice for YouTube: What You Must Disclose Before You Publish.

What we could not verify

Honesty about gaps beats a page that looks complete. Standing limits on everything above, as of September 3, 2026:

• How Murf sounds. Our listening benchmark is designed but not running. Nothing on this page or on any page it links to ranks Murf on audio quality, naturalness or speed.
• Anything Murf does not publish. Every page indexed here is built from the vendor’s own documentation. Where that documentation is silent, we record the silence rather than estimating around it.
• Whether a documented feature works as documented. We read documentation; we did not create an account or make an API call.
• Currency. Each linked page carries its own check date. This index inherits them and does not re-verify on load.

Change log

September 3, 2026 — First publication. Assembled from pages already published on this site; no new vendor claim was introduced by this page.

Published September 3, 2026 · This page shows no audio and reports no listening results, because our benchmark has not run. Corrections are recorded with their date on the corrections page. Independence note: this page contains no affiliate links, and no vendor paid for placement or influenced the order — see how we make money and our editorial policy.