Descript pricing

About

Descript is an AI-powered audio and video editing platform built around a document-style editor: you edit recordings by editing their transcript. The product line spans transcription, video and podcast editing, screen and remote recording, AI voice cloning, dubbing/translation, AI avatars, and “Underlord,” its AI video co-editor. Descript markets itself to individual creators, podcasters, marketers, and collaborative teams, and states that more than 6 million creators and teams use the product.

Commercially, Descript sells per-person subscriptions in four published tiers (Free, Hobbyist, Creator, Business) plus a sales-led Enterprise plan. The defining mechanic is a metered allowance baked into every paid seat: each plan includes a fixed number of media hours (uploaded/recorded media) and AI credits (consumption of AI features) per month, with paid top-ups available when the bundled allowance runs out. This makes Descript a seat-priced subscription with usage-based consumption layered on top.

Descript was founded in 2017 by Andrew Mason, the former CEO of Groupon, and is privately held. It has raised roughly $100M total from Andreessen Horowitz, Redpoint, Spark Capital and others; the headline round was a $50M Series C in November 2022 led by the OpenAI Startup Fund at a valuation above $500M (reported around $550M). That OpenAI relationship matters to the pricing story: the funding underwrote the AI-feature buildout — Underlord, generative video, AI Speech — that Descript later moved into its metered AI-credit pool. The company has also grown by acquisition, buying remote-recording service SquadCast in 2023 (now surfaced as the per-tier “Recording session hours” allowance) and the voice-cloning startup Lyrebird earlier in its history. For current rates, see the Descript pricing page.

Pricing summary : per-seat subscription with metered media hours and AI credits

Descript prices per person per month in USD. Each paid tier (Hobbyist, Creator, Business) bundles a monthly allowance of two metered dimensions — media hours and AI credits — and lets you buy top-ups when you exhaust either. Billing cadence is a toggle: annual billing is charged per month at a lower effective rate (up to ~35% off) and also grants bonus media hours and bonus AI credits, while monthly billing is higher and carries no bonus. This places Descript in the hybrid pricing model camp — seat-based access with credit-based billing for consumption.

The dimensions that move a Descript bill:

• Seats — $16/$24/$50 per person/month (annual) or $24/$35/$65 (monthly) for Hobbyist/Creator/Business; Free is $0. Creator scales to a team of 3 and Business to a team of 5, each seat billed separately.
• Media hours — included allowance per seat per month (10 / 30 / 40 hrs on paid tiers; 1 hr on Free); annual adds +5 (Creator) / +10 (Business) bonus hours.
• AI credits — included allowance per seat per month (400 / 800 / 1500 on paid tiers; 100 one-time on Free); annual adds +500 (Creator) / +1000 (Business) bonus credits.
• Top-ups — paid add-ons to buy more media hours or AI credits beyond the bundled allowance (available on paid tiers).
• Enterprise — custom-quoted, with custom AI credits, custom media minutes, SSO/SCIM, and flexible billing (including ACH).

What makes this different: the two metered allowances are denominated in user-legible units (hours of media, AI credits) rather than tokens, and annual commitment is incentivised not just with a price cut but with a larger usage allowance — a packaging lever most seat-priced tools don’t pull.

Pricing by product

Descript (Individual & team plans)

Usage top-ups & consumption units

Sales motions across products: PLG / self-serve for Free, Hobbyist, Creator, and Business (credit-card checkout, no sales contact); sales-led for Enterprise (custom quote, ACH/invoicing, SSO/SCIM).

Hidden costs : where the media-hour and AI-credit pools push the bill up

The sticker price on a Descript seat is not the whole story. Two metered pools — media hours and AI credits — sit on top of every seat, and the September 2025 change to counting every uploaded file as media minutes means the pool drains faster than most users expect. Below are two representative archetypes.

Archetype 1 — a solo podcaster on Creator who records multitrack. Creator includes 30 media hours/month, but a one-hour interview uploaded as separate host + guest tracks plus a cleaned mix can consume ~300 media minutes (5 hours) instead of 60 — so a few episodes a week can exhaust the bundle and force a media-minute top-up.

For a creator who budgeted “$24/month,” a busy production cycle can land near $140 — the top-ups, not the seat, dominate the bill.

Archetype 2 — a 5-seat Business team producing weekly video. Business includes 40 media hours and 1,500 AI credits per editor per month. A team running heavy AI features (dubbing at ~15 credits/minute, Studio Sound and Eye Contact at ~10 credits/use) and uploading raw multicam footage can blow past both pools.

So a “$250/month” team can reach ~$600 in a heavy month — the consumption layer can rival or exceed the seat layer, which is exactly the bill-shock pattern long-tenured users reported after the 2025 overhaul. Top-up prices are sourced from independent teardowns of the new model and may change; treat them as directional. See bill shock and cost unpredictability for why metered pools surprise buyers.

Want to estimate your own Descript bill? Use the Descript pricing calculator to model your monthly cost based on seats, media hours, and AI-credit consumption.

Pricing evolution : from transcription hours to media minutes and AI credits

Descript’s pricing has moved through three recognisable phases: an early transcription-and-Overdub subscription, a transcription-hour seat model after the OpenAI-backed AI relaunch, and the current media-minutes-plus-AI-credits hybrid introduced in September 2025.

Cadence

Tracked range: 2022 Q1–2026 Q2. Captured pricing snapshots show the headline seat prices held at $0 / $12 / $24 / Custom across every month from 2022-01 through 2024-05, so the transcription-hour era was price-stable; the move to $16 / $24 / $50 seats and the media-minutes + AI-credits rework arrived with the September 2025 overhaul. Dated funding and acquisition events are corroborated by Descript’s own announcements/help center and independent press coverage.

Notable changes

• 2022-11-15 — OpenAI Startup Fund leads a $50M Series C at a reported ~$550M valuation; Descript relaunches around AI editing (per The Information and TechCrunch coverage and Descript’s own “all-new Descript” announcement).
• 2023-08-15 — Descript acquires remote-recording service SquadCast (Bloomberg; Descript announcement), bundling remote recording into subscriptions.
• 2025-09-23 — Transcription-hour plans replaced by media minutes + AI credits with purchasable top-ups; existing subscribers moved to Legacy/Sunset plans ahead of auto-migration with notice (Descript’s “Understanding Your Legacy and Sunset Plan” help-center article, corroborated by multiple independent teardowns).

The September 2025 overhaul in detail

The 23 September 2025 change is the defining inflection in Descript’s pricing. Three things changed at once:

1. Metering unit swapped. The old paid tiers were sized by transcription hours (Hobbyist 10h, Creator 30h, Business 40h). The new tiers are sized by media minutes — and Descript now counts every uploaded or recorded file, transcribed or not. Uploading multitrack audio plus video for one piece of content multiplies consumption, so the same workflow that fit inside an old plan can exhaust the new pool.
2. AI features became a metered line item. Capabilities that were effectively uncapped inside a tier — Underlord, Studio Sound, and Overdub/AI Speech — moved into the AI-credit pool, with documented per-action costs (e.g. Studio Sound and Eye Contact ~10 credits/use, dubbing ~15 credits/minute, text-to-speech ~5 credits/minute).
3. Top-ups and a Legacy/Sunset migration. Purchasable top-ups for both pools arrived, and existing subscribers were placed on Legacy/Sunset plans that would auto-migrate to the new structure after notice. Some long-tenured creators publicly reported their effective bill jumping from a steady ~$30/month into the hundreds during heavy production cycles once the media pool was exceeded.

What’s unique : two human-legible usage pools on top of a seat

1. Two metered pools denominated in user-legible units. Most AI tools meter in tokens, which buyers can’t reason about. Descript meters in media hours (hours of audio/video) and AI credits with published per-action costs (“Studio Sound = ~10 credits”). A podcaster can map “an episode” to a known media-minute draw, which makes the consumption layer of this hybrid pricing model far more intuitive than a token meter — even if the per-file counting rule still surprises people.

2. Annual commitment buys allowance, not just discount. Descript’s annual toggle does two things at once: it cuts the effective monthly seat price by up to 35%, and it grants bonus media hours and bonus AI credits (Creator +5h/+500 credits; Business +10h/+1,000 credits). Pairing a price cut with a larger usage envelope is a packaging lever most seat-priced SaaS never pulls — it pushes heavy users toward commitment precisely because they’re the ones who need the extra allowance.

3. Seats and usage are deliberately decoupled. A Creator plan “scales to a team of 3” and Business “to a team of 5,” but each seat is billed separately and carries its own media-hour and AI-credit allowance (allowances are per editor). That keeps the seat-plus-usage pricing math additive and predictable per head, rather than pooling consumption across a shared team bucket the way some credit-based billing tools do.

4. Top-ups instead of forced upgrades. When a paid user exhausts either pool, Descript sells a top-up rather than forcing a tier jump. This is friendlier than a hard cap, but it also converts a fixed subscription into a variable bill — the mechanism behind the post-2025 bill-shock complaints.

5. The Free tier is a genuine on-ramp, not a demo. Free is $0 with no credit card, real (if watermarked) export, and a one-time 100 AI credits — enough to actually finish a small project. That low-friction entry is the PLG engine feeding the paid tiers.

Strengths & weaknesses

Billing UX : cadence toggle, allowance metering, and self-serve cancellation

• Monthly / Annual toggle — the pricing page exposes a billing-cadence switch; annual is charged per month at a lower effective rate and visibly states “Save up to 35% with annual billing” plus the bonus media hours / AI credits each annual plan adds.
• Media hours & AI credits metering — both metered allowances are denominated per editor per month (e.g. “30 media hours / month”, “800 AI credits / month”), with annual bonus amounts shown inline as chips (“+5 bonus hours”, “+500 bonus credits”).
• Top-ups — paid tiers advertise “Access to top ups for more media hours and AI credits,” letting users buy beyond the bundled allowance rather than forcing an upgrade.
• Automatic card billing — Hobbyist, Creator, and Business subscriptions are charged automatically to the saved credit card on the selected billing cycle (no invoicing on self-serve tiers).
• Self-serve cancellation — users can cancel from their Descript subscription page and can also delete their account and wipe their data; no sales contact required.
• Enterprise invoicing — custom invoicing options such as ACH transfer are available only via the Enterprise team, alongside flexible licensing and billing.
• Free-tier transparency — the Free plan requires no credit card to start, and the page includes a full “Compare all features and plans” matrix plus an FAQ defining media hours and AI credits.
• Special rates — students, educators, and non-profits can request special pricing via a form linked from the pricing FAQ.

Strategic wins : where Descript’s pricing decisions paid off

1. Metering in media hours and AI credits, not tokens

By denominating its usage layer in hours of media and credit-per-action, Descript made an AI-heavy product legible to non-technical creators. A podcaster can reason about “an episode” against a media-minute allowance in a way they never could against a token meter. This is the right call for a creator audience and a textbook application of choosing the right usage metric — the metric should map to a unit the buyer already counts. It also positions Descript well as AI editing commoditises and per-token costs fall.

2. Tying annual commitment to allowance, not just price

Descript’s annual plans deliver a discount and extra media hours plus AI credits. That double incentive targets the exact segment most likely to churn on a variable bill — heavy users — and converts them to a committed, higher-allowance plan. It’s a sharper version of the standard annual-discount play and a good example of packaging as a retention lever, complementing the seat model described in how AI companies structure pricing.

3. A Free tier that actually finishes work

Keeping a no-credit-card Free tier with real export and a starter credit grant gives Descript a genuine PLG funnel rather than a crippled demo. Creators can complete a small project, feel the product’s value, then hit the watermark/allowance ceiling — a natural upgrade trigger. This is the usage-based pricing on-ramp done well.

4. Top-ups absorb spikes without forcing tier jumps

Selling top-ups when a pool runs dry lets a Hobbyist or Creator handle an occasional busy month without permanently upgrading. Done transparently, this captures upside revenue from spikes while keeping the base subscription affordable — aligning vendor and buyer the way good outcome-based pricing trends do, since the user only pays more when they produce more.

Areas to improve : where the model creates avoidable friction

1. Make the “every file counts” rule visible before the bill

The single biggest source of post-2025 frustration is that media minutes are consumed per uploaded file, so a multitrack workflow can draw 4–5× the allowance of a single mixed file — a rule most users only discover after a top-up prompt. A real fix: surface a live “this upload will use ~X media minutes” estimate at the point of upload, plus a monthly burn-down meter. Pre-empting the surprise is the proven antidote to bill shock and cost unpredictability.

2. Offer team-pooled allowances on Business

Because allowances are strictly per-editor, a 5-person Business team can have one editor hitting top-ups while four waste most of their pool. A pooled-allowance option (sum the team’s media hours and AI credits into one shared bucket) would smooth this and reduce avoidable top-up spend — a packaging refinement consistent with usage-based pricing best practice for collaborative tools.

3. Smooth the Legacy/Sunset transition with clearer side-by-side math

The migration to Legacy/Sunset plans left existing customers comparing an old transcription-hour plan against a new media-minute plan with no easy translation. A published conversion guide (“your old 30 transcription hours ≈ N media minutes for a typical podcast workflow”) plus a personalised “your last 3 months under the new model would have cost…” estimate would defuse the migration anxiety that drove much of the community backlash.

4. Publish top-up pricing on the pricing page itself

Top-up prices currently surface mostly through third-party teardowns rather than the public pricing page. Putting the media-minute and AI-credit top-up rates directly on descript.com/pricing — the way the seat prices already are — would restore the transparency the rest of the page exemplifies and let buyers model a realistic worst-case bill up front.

Key takeaways

1. Pick a usage metric your buyer already counts. Media hours and AI credits work for creators because they map to “an episode” or “a Studio Sound pass” — units the customer can reason about. The legibility of the metric matters more than its technical precision.
2. Packaging beats discounting. Tying annual commitment to a larger allowance (not just a lower price) is a stronger retention lever for heavy users than a flat percentage off, because it solves their actual problem — running out of capacity.
3. Changing the metering unit is the riskiest pricing move you can make. Descript’s September 2025 switch from transcription hours to per-file media minutes was logically defensible but operationally jarring; the “every file counts” rule re-priced existing workflows overnight and produced real backlash.
4. Top-ups are friendlier than hard caps — but only if they’re visible. Letting users buy overflow capacity beats blocking them, yet it quietly converts a fixed subscription into a variable bill. Surfacing burn-down and estimated cost is what separates “flexible” from “bill shock.”
5. A real Free tier is a funnel, not a cost. Descript’s no-credit-card Free plan with genuine export and a starter credit grant is the engine that feeds paid conversion — the watermark and allowance ceilings are the upgrade triggers.

UBP implications

1. Hybrid seat-plus-usage is the default for creator and prosumer AI tools. Descript shows the pattern works outside developer infra: a predictable per-seat base plus a metered consumption layer captures upside from power users without pricing out the long tail. Expect more creative-software vendors to copy the two-pool structure.
2. The unit of metering is a product decision, not just a finance one. When Descript moved to counting every uploaded file, it changed how users work (encouraging single-mix uploads). Usage-based pricing teams must model the behavioural second-order effects of a metering unit, not only its revenue impact.
3. Migration design is half of any re-pricing. The Legacy/Sunset rollout proves that the transition mechanics — notice, grandfathering, side-by-side cost translation — drive customer sentiment as much as the new prices themselves. A defensible new model can still generate backlash if the migration leaves customers unable to translate their old bill into the new one.

Sources

• Descript pricing page (accessed 2026-05-31)
• Descript “Understanding Your Legacy and Sunset Plan” help article (accessed 2026-05-31)
• Descript “Track and understand your media minutes and AI credits” help article (accessed 2026-05-31)
• Descript blog (accessed 2026-05-31)
• Descript changelog (accessed 2026-05-31)

Bottom line

Descript is a per-seat creator-editing subscription with a genuinely usage-based core: every paid seat bundles metered media hours and AI credits, and paid top-ups absorb the overflow. The September 2025 shift from transcription hours to per-file media minutes plus AI credits made the model more legible on paper but re-priced existing workflows overnight — a reminder that for usage-based pricing, the choice of metering unit and the design of the migration matter as much as the headline price.

Want to compare Descript against other hybrid seat-plus-usage companies? Browse the pricing blueprint.

Pricing timeline : Major events on a vertical axis

Each milestone below corresponds to a public pricing change, product launch, or material adjustment. Major events use a filled marker; minor adjustments use a faded one.

Media minutes + AI credits replace transcription hours

Sep 2025

Descript overhauled its plans: transcription-hour allowances (Hobbyist 10h / Creator 30h / Business 40h) were replaced by two metered pools — media minutes (every uploaded/recorded file) and AI credits (AI-feature consumption) — and purchasable top-ups were introduced. Features that were previously uncapped inside a tier (Underlord, Studio Sound, Overdub/AI Speech) moved into the AI-credit pool. Existing subscribers were moved to Legacy/Sunset plans pending auto-migration with notice. Corroborated by Descript's Legacy/Sunset help-center article and multiple independent teardowns; some long-tenured users reported bill increases once media-minute pools were exceeded.

Feature-matrix pricing page; seats steady at $0 / $12 / $24 / Custom

May 2024

By mid-2024 the pricing page led with a full feature-comparison matrix above the four columns, but the displayed seat prices were unchanged from 2022: Free $0, a $12 tier, a highlighted $24 tier, and a Custom/Enterprise tier. Across every monthly Wayback snapshot from 2022-01 to 2024-05 the headline seat prices held at $0 / $12 / $24 / Custom — the transcription-hour era was price-stable; the later jump to $16 / $24 / $50 and the media-minutes + AI-credits rework did not appear until the September 2025 overhaul (after this archive window).

SquadCast acquisition adds remote recording

Aug 2023

Descript acquired remote-recording studio SquadCast (founded 2016). Remote recording was folded into Descript subscriptions, later surfacing as the per-tier 'Recording session hours' allowance (2h Free / 5h / 15h / 25h) on the current pricing matrix. Reported by Bloomberg and Descript's own announcement.

Pricing page redesigned to a feature-comparison matrix

Feb 2023

The pricing page moved from a four-card layout to a long feature-comparison table with the four plan columns (Free $0, $12, $24, Custom) pinned at the top. Per-tier prices were unchanged — $0 / $12 / $24 / Custom, matching the 2022 cards — so this was a presentation change rather than a repricing. The $24 column remained the highlighted/popular tier.

OpenAI Startup Fund leads $50M Series C

Nov 2022

OpenAI's Startup Fund led a $50M Series C at a valuation above $500M (reported ~$550M), with a16z, Spark, Redpoint and Daniel Gross participating, alongside an 'all-new Descript' product relaunch centred on AI editing. This funding underwrote the AI-feature buildout (Underlord, generative media) later metered as AI credits.

Archived baseline: four-tier card layout ($0 / $12 / $24 / Custom)

Jan 2022

Earliest captured Wayback snapshot. Descript showed a four-card 'Choose your plan' layout: Free $0, a $12 tier, a highlighted $24 tier, and a Custom/Enterprise tier ('Invite your whole team. Only editors pay'). These transcription-hour-era seat prices stayed constant through every monthly snapshot up to 2024-05.