what sign up
nixm nixm

← learn

How to make a picture talk with AI: the three ways, and a five-minute walkthrough

November 4, 2026

Pick by what you want the picture to do. For a face reading a script straight to camera, use a talking-photo avatar tool (HeyGen photo avatar, D-ID, A2E); monthly plans, minutes or credit caps, strong lip sync. For a short line said in a real scene with sound, upload the photo to nixm and say "animate this" with the line: the photo stays the frame, 8 seconds with sound, one credit, no plan; or cast the person into a new scene for a couple of credits. For a family photo that should move and breathe rather than speak, an animator like MyHeritage. The walkthrough below is the second route.

The three ways, and which one you want

Talking-photo avatars. Upload a face, paste a script, pick a voice; the tool animates the mouth to match for as long as the script runs. HeyGen, D-ID, Hedra and A2E do this. Best lip sync of the three routes, longest speech, and a monthly plan with a minutes or credit allowance; free tiers are watermarked. Right for a presenter, a message to camera, a joke with a historical portrait.

Animators. MyHeritage and similar add motion to an old photo: a turn of the head, a blink, a smile. Some add a generic voice. Moving, not talking. Right for a family photo you want to see breathe.

Generated video. Two routes, and the words you use pick between them. Say animate this and give the line: the photo becomes the first frame and stays the frame, the model adds one camera move, ambient action and sound, and a line in the prompt can be spoken; 8 seconds, one credit. Or cast the person as a character and put them somewhere new saying the line: nixm proposes the scene, you say go, and it renders a cast scene at one credit per five seconds. Shorter speech than an avatar tool, no plan, and a scene instead of a face on a background. This is what nixm does.

Walkthrough: make a picture say a line, no subscription

  1. Choose the photo. One person, face visible, decent light. A front-facing or three-quarter view gives the model the most to work with.
  2. Open nixm and upload it. You can plan in words without an account; rendering takes a free sign-in with no card.
  3. Write the line, short, and say when it comes. Eight seconds holds about fifteen spoken words. "I told you we should have left earlier" works; a paragraph does not. Ask for the line early, or the clip may wait before speaking.
  4. Say "animate this" with the line and the mood. That keeps the photo as the frame: one camera move, the room's sound, the line. One video credit, 8 seconds, 1080p. Do not name the person as a character here; that routes to a scene.
  5. Or cast them into a scene instead. If you want them somewhere new, doing something, say so: nixm casts the person from the photo, proposes the scene, and renders when you say go, at one credit per five seconds (ten seconds is two). The likeness holds; the picture around them is new.
  6. Add a shot if the line needs more room: 2 to 8 takes stitch into one video of up to 120 seconds.

The line is generated as dialogue with the picture. Lip sync works: a line you write is spoken on the character's lips. Short lines fit an 8-second shot; the lip-sync test shows one line that works and one that does not. Your first HD video with sound is free on a new account. No transcript of what you type is kept, and the uploaded photo dissolves with the signal after seven days.

What one render actually looks like

We ran the scene route once, on a public-domain portrait of Aletta Jacobs photographed around 1900, with the line "I told you we should have left earlier" and the setting "a quiet train platform at dusk, steam."

A composed still: a woman with curled grey hair, a high white collar and an embroidered bodice stands on a station platform at dusk beside a steam locomotive and a gas lamp, in sepia.
The composed still: nixm kept her likeness (hair, high collar, embroidered bodice) and built a platform and a locomotive around her from one sentence. A still costs two chat credits.
A frame from the rendered scene at six seconds: the same woman on a platform at dusk under a lamp, steam behind her, mouth open mid-line.
The scene at six seconds, mid-line. Her mouth is closed for the first five seconds, then moves with the voice for about two. Ask for the line early.

Two honest notes. The scene was recomposed, not a copy of the still: same person, same platform, different light. And it cost two video credits, because a ten-second cast scene is two; the animate-this route on the same photo would have been one credit and eight seconds with the frame unchanged. Photo: Max Büttinghausen, Groninger Archieven, public domain, via Wikimedia Commons.

How long can the picture talk?

Avatar tools: as long as your plan's minutes allow, in one take. Generated video: about fifteen words in an 8-second animate-this clip, about twenty in a ten-second scene, more by adding shots. If the point is a two-minute speech, an avatar tool is the right one. If the point is one line that lands, a scene beats a face on a background.

What it costs

Avatar tools sell monthly plans by minutes or credits; the entry tiers are compared in the HeyGen alternative page and what A2E's free credits buy. Animators are free with a watermark or a subscription. nixm is one video credit for an 8-second animate-this clip, or one credit per five seconds for a cast scene, bought once, never expiring, no plan.

Quick pick

  • A face reading a script, minutes long: HeyGen, D-ID or A2E.
  • A family photo that moves: MyHeritage or similar.
  • One line, said in a scene, with sound, no plan: nixm.

For an old family photo specifically, with sound and no subscription, see animate old photos; for a greeting, a video greeting from a photo.

Try one line

Upload the photo in the cinematic studio and write the line. After a free sign-in, say animate this, and look at the clip before you add another shot.

frequently asked

How do I make a picture talk for free?

Avatar tools have watermarked free tiers. On nixm you upload the photo and plan without an account; rendering takes a free sign-in with no card and one credit for an 8-second clip.

What is the best app to make a photo talk?

For a long script to camera, HeyGen, D-ID or A2E. For a short line said in a scene with sound and no subscription, nixm. For motion without speech, an animator like MyHeritage.

How long can the talking be?

Avatar tools run as long as your minutes allow. On nixm, about fifteen words in an 8-second animate-this clip or about twenty in a ten-second scene; 2 to 8 takes stitch into one video of up to 120 seconds.

Does the voice match the person?

Avatar tools let you pick or clone a voice. On nixm the voice is generated to fit the character and the line; you describe the mood and the delivery.

Is my photo kept?

On nixm it lives in the signal for seven days and then dissolves. Avatar tools store uploads on your account; check their retention terms.

Can I make a historical portrait talk?

Yes, with any of the three routes. Avatar tools give the longest speech; nixm puts the portrait in a scene saying one line with sound.

cast one character. put it anywhere.

try it now · cinematic studio

cast one character. put it anywhere.

cinematic ai video, with sound. nixm proposes every scene before a credit is spent.

start directing → first video free · no subscription

Enter your email and receive a link to sign in. No passwords, just a secure signal to connect.

why connect

  • 20 free credits to start
  • your signals, remembered
  • buy more credits anytime
  • image & video generation features
or