D-ID AI Avatar Generator: Pricing, Credits & Review

D-ID is an AI avatar and video platform for turning a photo, recorded presenter, script, or audio track into a talking digital-human video. Its web-based Creative Reality Studio is aimed at no-code creation, while its APIs support product integrations, batch rendering, video translation, and real-time visual agents.

The platform is best known as a talking photo AI and AI avatar generator, but its scope now includes expressive V4 avatars, custom avatars built from video, multilingual text-to-speech, voice cloning, translated lip sync, and interactive agents with custom knowledge.

What can D-ID create?

• Talking photos and avatar videos: V2 Photo Avatars animate a single frontal image from text or audio. V3 Instant and Pro Avatars use recorded footage for more natural movement, while V4 adds selectable sentiments and more expressive delivery.
• Multilingual video: D-ID says video creation and real-time interactions support 120+ languages. Studio can use synthetic voices, uploaded audio, or a cloned voice.
• AI video translation: Upload a talking-head video to transcribe, translate, clone the speaker's voice, and adapt lip movement. The official help guide lists 29 target languages.
• Visual AI agents: Combine an avatar, voice, LLM instructions, and a knowledge base for real-time WebRTC conversations that can be embedded in a website.
• Studio and API workflows: Non-technical users can build scenes in the editor; developers can generate offline videos, stream agents, manage avatars and voices, or store outputs in their own S3 bucket.

Explore D-ID's avatar models · Read the API quickstart

D-ID pricing and credits

D-ID is a paid subscription product with time-limited trials rather than a permanent production-ready free plan. Prices below were checked on August 20, 2026 and can change; confirm the currency, billing period, tax, and included usage at checkout.

Studio plans

The current Studio lineup is Trial, Lite, Pro, Advanced, and Enterprise. D-ID's official material states that paid Studio plans start at $4.70/month for Lite on annual billing and reach $108/month for Advanced; Enterprise is quoted. Annual subscriptions are charged for the year, while credits are issued monthly.

• The Studio trial lasts 14 days.
• Trial and Lite output carries D-ID branding and is licensed for personal use only.
• Voice cloning is available on Pro and above. Pro includes one cloned voice; Advanced includes three.
• Pro and Advanced include commercial-use rights, but generated output still carries an AI mark or custom logo depending on plan.

Check live Studio pricing · Understand subscription billing

API plans

The public API pricing page currently lists:

• Trial — $0 for 14 days: up to 3 minutes of offline video and 10 minutes of streaming, a full-screen watermark, and a personal-use license.
• Build — $14.40/month billed annually: 64 credits, up to 16 offline-video minutes or 32 streaming minutes, D-ID watermark, and personal-use licensing.
• Launch — from $35/month billed annually: the base 180-credit tier provides up to 45 offline-video minutes or 90 streaming minutes, plus commercial licensing, premium voices, and one voice clone.
• Scale — from $138.60/month billed annually: the shown 800-credit tier provides up to 200 offline-video minutes or 400 streaming minutes, with higher usage selections available.
• Enterprise — custom quote: intended for larger usage, support, security, branding, and deployment requirements.

One credit covers up to 15 seconds of offline video. Streaming consumes 0.5 credit per up-to-15-second interval. Every started interval rounds up, so a 40-second offline video costs three credits. Credits reset monthly, do not roll over, and unused credits expire.

View current API pricing · See the credit calculation

Commercial use and consent

D-ID's license depends on the plan. Trial, Lite, and API Build are personal-use only; their videos may be shared socially but cannot be used commercially. Pro, Advanced, Enterprise, API Launch, and Scale permit commercial use, including ads, marketing, and business content, provided you own or have permission to use every uploaded image, voice recording, video, and script.

Custom avatars and cloned voices require consent. V3 avatar creation records a webcam consent statement and checks the face and, in some flows, the voice against the source footage. Voice cloning requires the speaker to pronounce a consent statement; cloning someone else's voice is allowed only with that person's explicit permission.

Review commercial-use rights · Review avatar consent

Privacy and biometric data

Creating an avatar, animating a face, cloning a voice, or translating a video can involve biometric processing such as face geometry, voiceprints, gait, or other features. D-ID says biometric information is not used to improve its products or train the models that power them, but it may be shared with service providers for identity verification or moderation. Voice samples and templates are stored when needed for future cloned-voice use.

Retention depends on the product and storage choice. According to D-ID's privacy policy, ordinary API uploads are automatically erased within 24 hours after completion if the user does not delete them first; Agent data is erased within 72 hours and Insight Service data within 14 days. Studio uploads persist until the user deletes them. API users can instead request persistence in encrypted D-ID storage or supply signed S3 URLs so media stays in their own bucket. D-ID warns that it does not provide backup beyond the stated retention rules.

Read the privacy policy · Read the biometric privacy policy

Important limitations

• Short clips are rounded up: each partial 15-second block consumes a full credit for offline video, and unused monthly credits do not roll over.
• Low-cost plans are not commercial licenses: Trial, Lite, and Build cannot be used for advertising, paid client work, or business content.
• Watermarks remain on most self-serve tiers: Trial has a full-screen mark, Lite and Build use D-ID branding, and Pro/Launch/Scale use an AI watermark. Removal or custom branding requires a higher plan.
• Translation favors a controlled talking head: D-ID recommends one front-facing speaker, clean audio, and a clearly visible mouth. The help guide caps an uploaded video at five minutes and 2 GB; if several people speak, one cloned voice is applied to the entire output.
• Moderation can reject inputs: a missing or obscured face, celebrity footage, third-party image screening, or flagged text/audio may prevent generation.
• Feature availability varies: V3 Instant Avatars are for offline generation and do not work with streaming Agents; V4 custom avatars are Enterprise-only. Proofreading translated scripts is currently described as an Enterprise feature.
• Sensitive biometric uploads need care: obtain explicit permission, understand retention, and avoid uploading third-party faces or voices without the necessary rights.

Who is D-ID best for?

D-ID is a strong fit for teams that need short talking-avatar videos, localized training or marketing clips, or a visual agent embedded in an application. The API is particularly relevant when a product needs photo-to-video or real-time avatar output. It is less suitable for buyers seeking an unrestricted free tier, watermark-free entry pricing, multi-speaker translation, long-form flat-rate video, or an all-purpose timeline editor.

Frequently asked questions

Is D-ID free?

D-ID offers a 14-day trial, not a permanent production-ready free plan. The API Trial is $0 for 14 days and includes up to three offline-video minutes and ten streaming minutes. Trial output is watermarked and licensed for personal use only.

How many D-ID credits does a video use?

Offline video costs one credit for each started 15-second interval. A 40-second video therefore uses three credits. Streaming uses half a credit per started 15-second interval. Credits refresh monthly and do not roll over.

Can I use D-ID videos commercially?

Yes, on Pro, Advanced, Enterprise, API Launch, or API Scale, assuming you have rights to every input. Trial, Lite, and API Build are limited to personal use.

Can D-ID translate a video and keep the speaker's voice?

Yes. Video Translate transcribes and translates the speech, clones the speaker's voice, and adjusts lip movement. It works best with one clearly visible, front-facing speaker and clean audio; multi-speaker videos receive a single cloned voice.

Does D-ID train on my face or voice?

D-ID's biometric policy says biometric information is not used for product improvement or AI model training. The service still processes and may retain biometric data to create and maintain avatars or cloned voices, verify identity, prevent fraud, and moderate content, so review both privacy policies before uploading sensitive media.