LetzAI
Canvas for AI filmmakers: storyboard, generate multi-model shots, cut and share previews.
LetzAI is a browser-based creative workspace that aggregates image and video models from several providers (including Seedance, Kling, Veo, Flux and Nano Banana) on one infinite canvas. Since its Canvas launch in June 2026, an AI assistant can plan storyboards, generate stills shot by shot and animate panels into video, with a Present mode for shot-by-shot playback and export. It is aimed at previz, pitch work and agency client review.
- Freemium pricing
- Easy learning curve
What LetzAI can do
- Infinite canvas Yes
- AI agent Yes
- Node / flow builder Partial
- Public API Yes
- MCP server Not confirmed
17 of 26 capabilities confirmed
| Capability | Supported | Details |
|---|---|---|
| Text to image | Yes | |
| Image editing | Partial | Image edit is listed on the API integration pages; not confirmed on the product page |
| Text to video | Yes | Animate tab text-to-video since Feb 2026 (Kling V3, Veo 3.1, WAN 2.5) |
| Image to video | Yes | Animate any Canvas frame or storyboard panel; first and last frame control varies by model |
| Video to video | Partial | Video-editing models listed (Beeble SwitchX, Kling O3 Pro Edit, Veo 3.1 Extend Video); no general restyle mode confirmed |
| Native audio in video | Partial | Veo 3.1 native audio; depends on the chosen model |
| Lip-sync | Not confirmed | |
| Voice / TTS | Not confirmed | |
| Music & SFX | Not confirmed | |
| Upscaling | Yes | Magnific Creative and Precision V2 (2x to 16x), Nano Banana Upscaler (up to 4K), Seedance video upscalers |
| Character consistency | Not confirmed | |
| Camera control | Partial | Shot-by-shot storyboard generation and Present-mode sequencing; explicit camera-move controls not verified |
| Custom training | Partial | Custom model training via the Model Training API, Enterprise tiers only |
| Third-party models | Yes | Aggregates third-party models; see modelsOffered |
| Capability | Supported | Details |
|---|---|---|
| Infinite canvas | Yes | LetzAI Canvas (launched June 2026): infinite board for generation, storyboards and review |
| Node / flow builder | Partial | Board nodes the assistant can read and rearrange; not a visual flow or node-chain builder |
| AI agent | Yes | Canvas AI Assistant plans storyboards, generates stills, animates panels and runs multi-step jobs; reusable Skills and per-page Memory |
| Chat assistant | Yes | LetzAI Chat with text models such as Claude Sonnet 5 and GPT-5.5 |
| Team collaboration | Yes | Real-time shared Canvas, Owner/Editor/Viewer roles, org-wide sharing, client preview links |
| Capability | Supported | Details |
|---|---|---|
| Public API | Yes | Public API with documented endpoints and copyable model IDs; the pricing page and product page disagree on whether API access is included in every plan |
| MCP server | Not confirmed | |
| Mobile app | Not confirmed | |
| Desktop / local | No | Browser-based ('all in one browser tab') |
| Open source | No |
| Capability | Supported | Details |
|---|---|---|
| Free tier | Partial | Free account to browse and explore models; Canvas and the assistant require a paid subscription |
| Commercial use | Not confirmed |
Yes Partial No Not confirmed by the vendor
What makes LetzAI different
- 01
One workspace for storyboard, generation, shot-by-shot video animation and client previews, with no export step between tools
- 02
Multi-model access in one subscription: 40 active base models on the models page, including Seedance 2.5, Kling V3, Veo 3.1, Flux and Nano Banana
- 03
Client preview links that need no account, plus comments and colour labels for review rounds
- 04
Bring Your Own Keys on Pro and higher, so heavy users can route generation through their own provider accounts
- 05
Upscalers (Magnific Creative, Magnific Precision V2, Nano Banana Upscaler) built into the same catalog
About LetzAI
LetzAI started as a multi-model image generator and has grown into a production hub. Its catalog on the models page lists 40 active base models in October 2026, covering image models (Nano Banana 2.1, GPT Image 2.5, Seedream 5 Pro, Flux 2), video models (Seedance 2.5, Kling V3, Flux 3 Video, MiniMax H3 Max, Gemini Omni 1.1 Flash, Veo 3.1 variants), upscalers (Magnific Creative and Precision V2, Nano Banana Upscaler) and chat models. Pricing is per-credit, and the catalog states that availability and pricing may change.
What changed in 2026: text-to-video went live in February with Kling V3, Veo 3.1 (native audio) and WAN 2.5. Canvas launched on June 9, 2026 as an infinite board where you generate images and videos, lay out storyboard strips, animate panels, and reorder clips in Present mode before rendering. The Canvas AI Assistant reads the board and selection and can plan storyboards, generate stills, animate panels and pull web references onto the board. Team review (comments, labels, real-time cursors, preview links for clients without an account) came with the same workspace.
For a filmmaker, LetzAI is best used as the previz and review layer: block out a sequence on the canvas, test the same shot across several video models side by side, then hand the approved shots to an NLE. Its own timeline is for sequencing and quick cuts, not finishing. Custom model training and the Model Training API are limited to the Enterprise tiers, and BYOK is available from the Pro tier up.
What's new in LetzAI
-
Nano Banana 2.1 and 2.1 Upscaler added
Google image model and its upscaler (up to 4K) are listed on the LetzAI models page.
Source: letz.ai ↗ -
Claude Opus 5.5, Claude Fable 5.1, GPT Image 2.5 and Gemini 3.8 Flash listed
Chat and image models added to the catalog in September, shown with listed context and per-token prices where stated.
Source: letz.ai ↗ -
Seedance 2.5 and Flux 3 Video listed
ByteDance's Seedance 2.5 (up to 30s, 480p to 1080p) and Black Forest Labs' Flux 3 Video (up to 20s, HD or Full HD) joined the video catalog, alongside Gemini Omni 1.1 Flash (up to 4K).
Source: letz.ai ↗ -
Canvas launched
Infinite board for generating images and video, storyboarding, Present-mode shot timeline with 4K export, AI assistant and team review. Requires a paid subscription.
Source: letz.ai ↗ -
Text-to-video and model tagging
Text-to-video added with Kling V3, Google Veo 3.1 (native audio) and WAN 2.5, plus Magnific upscalers and a public models catalog with copyable API identifiers.
Source: letz.ai ↗
Models available in LetzAI
-
Seedance 2.5 ByteDance Up to 30s; 480p to 1080p
-
-
Kling V3 Kuaishou Up to 15s; multi-shot
-
FLUX 3 Video Black Forest Labs Up to 20s; HD or Full HD
-
MiniMax H3 Max MiniMax 5 to 15s; 480p or 768p
-
Gemini Omni 1.1 Flash Google Up to 4K
-
-
WAN 2.5 Alibaba Budget text-to-video option, Feb 2026
-
-
GPT Image 2.5 Sunburst OpenAI Up to 4K; listed Sep 2026
-
-
-
Grok Imagine 2.0 xAI 1K and 2K; $0.04 per image
-
Magnific Creative Magnific 2x to 16x
-
Key Features
- Canvas: infinite board for generating images and videos side by side, with comparison of variants
- Storyboard strips with shot-by-shot stills, and animation of storyboard panels into video
- Present mode: fullscreen shot timeline with split, trim and reorder, then render and export
- Canvas AI Assistant that reads the board and selection, plans storyboards and runs multi-step jobs
- Reusable assistant Skills and per-page Memory
- Real-time collaboration with live cursors, Owner/Editor/Viewer roles and org-wide sharing
- Comments, colour labels and preview links that clients can open without an account
- Text-to-video with Kling V3, Veo 3.1 and WAN 2.5, and image-to-video across the model catalog
- Magnific and Nano Banana upscalers up to 16x / 4K
- Models catalog with copyable identifiers for API use
- Bring Your Own Keys on Pro and higher
- Custom model training through the Model Training API on Enterprise tiers
When to reach for it — and when to skip
Reach for it when…
- Storyboard, generation, animation and client review sit in one browser workspace
- Many current video models (Seedance 2.5, Kling V3, Veo 3.1, Flux 3 Video, MiniMax H3 Max) on one subscription
- Client preview links without an account make agency approval rounds simple
- BYOK on Pro and above lets heavy users pay providers directly
Skip it when…
- Canvas and the assistant need a paid plan, and assistant chat draws from the credit balance
- Credits are non-refundable and generation prices can change without notice
- The model list changes often; per-generation credit costs are not on the plan pages
- The Present-mode timeline is for sequencing and quick cuts, not finishing edits
Best For
✓ Ideal for
- Storyboards and animatics for pitch decks, shot by shot
- Testing one shot across several video models on the same board before committing credits
- Agency client review with preview links and comments
- Animating storyboard panels into rough cuts for a director or producer to approve
✗ Not built for
- Finishing and colour work, which still belongs in Premiere Pro or DaVinci Resolve
- Lip sync, voice cloning and music or SFX, which are not offered in the current catalog
Working Tips from Filmmakers Using LetzAI
- 01 Block the sequence in Present mode before generating video; reordering clips is cheaper than regenerating shots
- 02 Test a shot on the cheapest video model first, then rerun only the keepers on a premium model
- 03 Use the assistant for repetitive board work (layout, storyboard strips), but check its credit cost in the balance first
- 04 Export stills and clips from Canvas and finish in an NLE, because Canvas is not a finishing tool
LetzAI pricing in 2026
| Plan | Price | What you get |
|---|---|---|
| Beginner | EUR 8.25/mo yearly (EUR 99/yr) |
|
| Fun | EUR 20.75/mo yearly (EUR 249/yr) |
|
| Pro | EUR 62.42/mo yearly (EUR 749/yr) |
|
| Ultra | EUR 165.83/mo yearly (EUR 1,990/yr) |
|
| Enterprise Basic | EUR 416.67/mo yearly (EUR 5,000/yr) |
|
| Enterprise Plus | EUR 1,250/mo yearly (EUR 15,000/yr) |
|
Prices checked against LetzAI's pricing page on 9 Oct 2026. Vendors change plans often, so confirm before you buy.
LetzAI vs Krea vs Higgsfield
| LetzAI | Krea | Higgsfield | |
|---|---|---|---|
| Price | From EUR 8.25/mo billed yearly (EUR 99/yr, Beginner); Canvas needs a paid plan | Basic $9/mo; Pro $35/mo ($21 yearly); Business $200/mo ($160 yearly); see the vendor's pricing page for other plans | Free tier; paid plans from €19/mo billed annually (prices exclude VAT) |
| Infinite canvas | Yes | Yes | Partial |
| Node / flow builder | Partial | Yes | Not confirmed |
| AI agent | Yes | Yes | Yes |
| Chat assistant | Yes | Yes | Yes |
| Team collaboration | Yes | Yes | Yes |
| Text to image | Yes | Yes | Yes |
| Image editing | Partial | Yes | Yes |
| Text to video | Yes | Yes | Yes |
| Image to video | Yes | Yes | Yes |
| Video to video | Partial | Partial | Yes |
| Native audio in video | Partial | Yes | Yes |
| Voice / TTS | Not confirmed | Yes | Yes |
| Music & SFX | Not confirmed | Yes | Yes |
| Upscaling | Yes | Yes | Yes |
| Character consistency | Not confirmed | Partial | Yes |
| Camera control | Partial | Not confirmed | Yes |
| Custom training | Partial | Yes | Not confirmed |
| Third-party models | Yes | Yes | Yes |
| Public API | Yes | Yes | Yes |
| MCP server | Not confirmed | Not confirmed | Yes |
| Mobile app | Not confirmed | Not confirmed | Partial |
| Open source | No | Partial | No |
| Free tier | Partial | Yes | Yes |
| Commercial use | Not confirmed | Yes | Not confirmed |
Each column comes from that tool's own researched page. "Not confirmed" means the vendor doesn't say.
Open the full LetzAI vs Krea vs Higgsfield comparison →Use Cases
Tags
Sources
Facts on this page were checked against LetzAI's own pricing, product and changelog pages on , then cross-checked with independent coverage where vendor pages were silent. Anything we couldn't confirm is marked "Not confirmed" rather than guessed.
- LetzAI pricing page letz.ai
- LetzAI models catalog letz.ai
- Introducing LetzAI Canvas (June 2026) letz.ai
- LetzAI February 2026 updates letz.ai
- LetzAI Canvas product page letz.ai
- LetzAI home page letz.ai
- Krea pricing page (comparison) krea.ai
Spotted something out of date? Tell us and we'll re-check it.
Discussion
No comments yet — be the first.