Changelog | Rewind.ai
Every feature shipped, model added and bug fixed, in order.
March 25, 2026Platform Launch: 400+ AI Tools
- 400+ free AI tools available at launch across chat, image generation, video creation, text to speech, speech to text, music, writing, coding, design and translation
- 400+ models included, covering self-hosted open-source models alongside models from OpenAI, Anthropic, Google and Meta
- Platform interface localized to 99 languages
March 25, 2026Chat History Sidebar
- Signed-in users now have a persistent chat history sidebar
- Switch between past conversations or start a new one without losing earlier context
March 25, 2026Auto Top-Up Feature
- Token balance can now refill automatically when it drops below a threshold you set
- Configure both the trigger threshold and refill amount in account settings
March 25, 2026Enterprise Plans
- Enterprise subscription plans now available, up to $4,999/mo
- Includes priority support, dedicated capacity and custom integrations
March 24, 2026Initial Launch
- Rewind.ai goes live with 400+ free AI tools accessible without an account
- Self-hosted GPU inference runs open-source models under Apache 2.0 and MIT licenses only
- Daily usage limits apply for anonymous and signed-in users; paid plans have no cap
FAQ
Updates ship on a rolling basis: new tools, model upgrades, bug fixes and performance improvements. Significant changes are recorded on this changelog page when they go out.
Rewind.ai launched with 400+ free AI tools covering chat, image generation, video creation, text-to-speech, speech-to-text, music generation, writing, coding, design and translation.
The sidebar shows previous conversations for signed-in users. It sits alongside the chat interface so you can switch between sessions or open a new one without losing earlier work.
When your token balance falls below a threshold you define, auto top-up purchases additional tokens automatically. You set both the trigger level and the refill amount in your account settings.
Most updates deploy with zero downtime via rolling restarts. If infrastructure work requires a maintenance window, it is scheduled during low-traffic periods and posted to the status page in advance.
The platform offers 400+ models. These include self-hosted open-source models on our GPU servers and models from providers such as OpenAI, Anthropic, Google and Meta.
Yes. Submit a request through the contact form. Requests are prioritized by user demand. The roadmap page shows what is planned and what is currently being built.
The platform interface is localized to 99 languages. The translation tool separately supports 450+ languages via the MadLAD-400 model.
Yes. Enterprise plans go up to $4,999 per month and include priority support, volume pricing and SLA-backed uptime.
All API endpoints maintain backward compatibility. If a breaking change becomes necessary, it will be versioned and announced with sufficient lead time so integrations can continue without interruption.
Rewind.ai launched on March 24, 2026 with self-hosted GPU inference, open-source models under Apache 2.0 and MIT licenses, daily free limits for all users and no cap for paid plans.
The roadmap page lists completed features, items currently in progress and what is planned next.