Flowith is an AI workspace built around an infinite branching canvas, where a single project can fork into multiple threads instead of scrolling through one long chat, and it gives users access to Claude, GPT, Gemini, DeepSeek and image or video models inside that same canvas. For Moroccan and diaspora founders juggling research, content drafts and visual assets across a lean team, it offers one subscription instead of several separate AI tools, though it works better for planning, drafting and comparing outputs than for coding or technical automation.
Key features
- Branching canvas: fork any prompt into separate threads to compare outputs side by side, instead of losing earlier attempts in a scrolled chat
- Access to 40+ AI models on paid plans, including Claude, GPT, Gemini and DeepSeek, plus dedicated image and video generation models
- Agent Neo, an autonomous agent mode built for longer, multi-step tasks such as research briefs and reports, working through steps inside the canvas rather than one exchange at a time
- Knowledge Garden, a workspace area for storing uploaded files and reference material so it stays available across projects
- Batch mode for image and video generation, running several variations at once on paid tiers
- Specialized modes for chat, image, video, slides and simple websites, so different output types stay inside one tool
- Commercial license for generated output included on paid plans
- Web app plus an iOS app for working outside the browser
- Concurrent task limits that scale by plan, from single-digit tasks on the free tier up to unlimited concurrency on the top plan
Pricing
Flowith runs on a credit system rather than flat seat pricing, and credits do not roll over between billing cycles. There are four tiers:
- Starter (free): $0/month, 300 credits as a one-time trial allowance, limited model access, no commercial license
- Pro: $19.90/month, or $17.91/month billed annually ($214.92/year), 20,000 credits a month, access to 40+ models, commercial license, 2x batch mode, up to 50 concurrent tasks
- Ultimate: $49.90/month, or $44.91/month billed annually ($538.92/year), 50,000 credits a month, an Unlimited Pack giving 365-day access to a set of top-tier models without spending credits on them, 4x batch mode, up to 100 concurrent tasks
- Infinite: $499.90/month, or $399.92/month billed annually ($4,799.04/year), 500,000 credits a month, the full Unlimited Pack across models, unlimited concurrent tasks, and priority support
Video generation uses far more credits than text or images, in some cases several thousand credits for one short clip, so teams planning to generate video regularly should budget for the Ultimate tier or higher rather than Pro.
Prices change over time. Figures above are based on Flowith’s official pricing page (Flowith.io/pricing).
Pros and cons
| Pros | Cons |
|---|---|
| Branching canvas keeps every draft and model comparison visible instead of buried in a long chat history | Credits do not roll over, and video generation can burn through a month’s allocation in a handful of clips |
| One subscription covers 40+ models (Claude, GPT, Gemini, DeepSeek, image and video generators) instead of paying for each separately | Independent review volume is still thin: a small number of G2 reviews and limited Reddit or Product Hunt discussion, so there is no broad consensus yet |
| Agent Neo can carry a research or reporting task through multiple steps without re-prompting at each stage | No terminal or file-system access, so it is weaker than dedicated coding agents for technical automation |
| Commercial license included on paid plans, relevant for agencies producing client deliverables | Limited input file-type support has come up repeatedly as a complaint from users |
| Batch mode lets you generate several image or video variations in one pass on paid tiers | The canvas does not save your position, so relocating a specific branch in a large project takes extra digging |
Who it’s best for
Flowith suits solo founders, small agencies and lean marketing or content teams in Morocco and the diaspora who are currently paying for several separate AI subscriptions (a chatbot, an image generator, maybe a research tool) and want them consolidated into one workspace with a visible history of every draft and version. It is also a reasonable fit for anyone producing content in more than one direction at once, such as comparing an Arabic, French and English version of the same pitch or social post, since the branching canvas is built for exactly that kind of side-by-side comparison.
It is a weaker fit for teams whose main need is software development or technical workflow automation, since Flowith has no terminal or file-system access and was not built as a coding agent. It also makes less sense for anyone who only needs occasional help from a single chatbot, since the free tier caps out quickly and the paid tiers are priced for regular, credit-heavy use rather than light, occasional questions. Given how thin independent reviews still are, teams that need proven track record and a large support community before committing budget should treat this as an early trial rather than a settled choice.
How to get started
- Create a free Starter account at Flowith.io to get 300 trial credits
- Open a project on the canvas and fork one prompt into two branches to see how the comparison view works
- Run one real multi-step task, such as a short research brief, through Agent Neo to judge how it handles a longer workflow
- If the canvas fits how your team already works, upgrade to Pro for full model access and the commercial license
FAQ
Is Flowith free to use? There is a free Starter plan with 300 one-time credits and limited model access, useful for testing the canvas but not licensed for commercial work.
What is Agent Neo? Agent Neo is Flowith’s autonomous agent mode, designed to work through longer, multi-step tasks like research and report writing inside the canvas rather than a single back-and-forth chat.
Does Flowith replace tools like ChatGPT or Claude? No. It sits on top of them, giving access to Claude, GPT, Gemini, DeepSeek and other models inside one branching workspace rather than replacing any single model.
Is Flowith suited to coding tasks? Not particularly. It has no terminal or file-system access, so teams needing coding automation are better served by agents built specifically for that purpose.
Disclosure: Moroccopreneur may earn a commission if you sign up for Flowith through the links in this article. This does not change what you pay, and it does not influence which tools we choose to cover.
