Your AI SaaS business, ready by this week!
Try demoAikeedo v5.0 Is Live: Canvas, Image & Video Templates, a Community Feed, and Translated Plugins
Published Sep 15, 2026 ⦁ 10 min read
Your AI SaaS business, ready by this week!
Try demoPublished Sep 15, 2026 ⦁ 10 min read
Aikeedo v5.0 is here — with Canvas documents in chat, clarifying questions, media generation from chat, a guided Writer, image and video templates, a community feed, a rebuilt interface, localization tooling, and translations for every plugin.
v4 made Aikeedo a complete business platform. v5 makes it feel like the AI products your users already love.
Chat now writes documents beside the conversation, asks before it guesses, and generates images, video and voice without leaving the thread. Imagine and Video get templates and a community feed. The whole interface has been rebuilt around a single sidebar. And every official plugin is now translated, with new tools to manage and complete translations yourself.
Aikeedo v5.0 is live.
Here is everything that is new.
Ask most chat tools for a report, a landing page or a script and they print the whole thing into the thread. Ask for a change, and they print it again. It is slow to read, hard to compare, and it burns tokens on every revision.
Canvas gives those documents their own place. When the assistant produces something that is really a document — a markdown report, a code file, an HTML page, an SVG — it writes it to a panel next to the chat. The conversation shows a compact card, and the document opens beside it.
Short standalone writing gets its own treatment. Emails, messages, social posts and letters render as a card directly in the conversation — and an email can be opened straight in Gmail, Outlook or the user's default mail app, subject and body prefilled.
Canvas started as a customer request — fewer rewrites, fewer tokens — and it brings the workflow ChatGPT and Claude made standard to your own platform.
Vague requests get vague answers. v5.0 lets the model do what a good colleague would: ask first.
With clarifying questions enabled, the model can pause its turn and ask a few multiple-choice questions — one at a time, with descriptions, multi-select where it makes sense, and a skip button. The user picks their answers, and the model continues with exactly the context it needed. The questions and answers stay in the thread as a compact summary.
It works across OpenAI, Anthropic, xAI, Cohere, Ollama and custom models, and you control it like any other capability: globally, per plan, and per user. The question turn and the answer turn are billed as two separate requests.
In-chat image generation has been rebuilt from the ground up — and widened well beyond images.
The chat can now generate images, videos, voice-overs and transcripts in the middle of a conversation. Results appear as inline cards instead of tiny links that opened a new tab, and long-running video generations resolve in place, even after a reload.
The most important change is editing. Before, asking to "make her jacket red" could only start over. Now the model can hand an earlier image — generated or uploaded — back to the provider as a reference, so adjustments actually adjust.
When the chat generates a voice-over, it shows a voice picker right in the conversation, so users can hear the options and choose a voice without leaving the thread.
All of it sits behind a single Media generation capability you can switch on per plan.
The Writer used to turn a filled-in form into a read-only document. If the result was not quite right, the only option was to fill in the form again. In v5.0 it becomes a front door to chat.
When a user opens a Writer template, a short wizard asks the template's questions one screen at a time and lets them pick a model. When they finish, the answers start a regular chat conversation — so the result can be refined with follow-ups, moved into Canvas, or extended with any tool your plan allows.
Two details matter for operators:
All 130+ templates carry over. Existing Writer documents are converted into conversations during the update, so nothing your users generated is lost.
Image and video generation got the biggest visual overhaul in this release.
Templates. You can now publish admin-authored templates for Imagine and Video, shown as cards with preview artwork and an optional hover animation. A template can pin settings such as aspect ratio or duration, require input images (think "turn your photo into a sticker pack"), and allow extra instructions. Lock a template and users get the result without ever seeing your prompt or model — which turns your best prompts into a product rather than something anyone can copy. v5.0 ships with 39 image and 24 video templates you can import in a click.
Community feed. Users can share their best generations to a feed visible to everyone on your platform, and start from what others have made. Locked-template generations never reveal their prompt. You control the feature, plus a minimum number of shared generations before the feed appears, so nobody lands on an empty page.
A new composer. The Imagine and Video composers now match Chat. The settings people use most sit right in the toolbar as chips that follow the selected model. Reference images can be tagged in the prompt with @image1. Enhance rewrites a rough prompt into a detailed one, and Surprise me writes a fresh prompt from scratch — both powered by a model you choose.
Reference images are now saved with every generation, shown in the details panel, and restored when a user hits Remix. A separate Edit button feeds the generated image straight back in as a reference.
v5.0 rebuilds the app shell around one idea: everything a user needs should be in the same place, in every tool.
Recents in the sidebar. Each tool's history moves out of a slide-out panel and into the left sidebar, scoped to the tool you are in. Rename, share, delete or open in a new tab right from the row. Status dots show at a glance what is queued, processing, failed or unread.
A head bar and Details panel. Every tool now shows the open item's title, a share button and its actions in one consistent bar. The new Details panel shows what the item cost, which model produced it, and insights that were recorded but never visible before — like a chat's credits broken down by model.
A proper Settings area. Account, security, capabilities, API keys, affiliates and workspace billing now live in one Settings section with its own navigation. Users can add personal instructions for chat, tell the AI what to call them, and revoke API keys.
Add a contextual New button under the logo, a single merged workspace and account menu, and redesigned dialogs that finally behave on phones, and the whole product feels noticeably more polished.
Aikeedo has shipped in 30 languages for a while. v5.0 makes localization something you can manage yourself, and extends it to every plugin.
All plugins are translated. Every official plugin, from Chatbots and Migration to payment gateways, tax engines and email marketing integrations, now ships its own translations. Update each plugin to get them.
Manage languages from the admin panel. Choosing which languages your users can pick is now a screen — Settings → Languages — with a default language and a completion percentage for each, instead of a JSON file on the server.
Translation tooling for operators. Customized a template, or want to finish a language yourself? New command-line tools extract every translatable string, translate the catalogs with an AI model of your choice (OpenAI or any OpenAI-compatible provider), and rebuild the browser dictionaries. Updates now keep your own translation edits instead of overwriting them.
Fewer English leftovers. Interface strings that could not be translated before — mostly messages shown by the browser, like toasts and notifications — are now part of the translation catalogs.
v5.0 brings the catalog up to date with the latest releases:
Models with announced shutdowns are retired, including Sora 2, GPT-4, the o-series, Claude Sonnet 4.0, Imagen 3 and 4, and Veo 2 and 3. We also fixed requests failing when thinking was turned off on models that cannot disable it, and Writer requests failing on newer models that reject the temperature setting.
A major version is also a chance to let go of what no longer earns its place.
Coder is gone. Chat already handles code well, and Canvas now gives code files a proper home. Existing code documents are converted into conversations.
Composer is gone. The music tool depended on a provider that discontinued its API. Existing compositions are deleted during the update, so download anything you want to keep first.
The Writer's form-and-document flow is gone, replaced by the guided wizard described above, along with the legacy completions API endpoint.
v5.0 is a major release. Back up your files and database first, then update as usual from the admin panel — database migrations run automatically. After the core update, update every installed plugin: plugins built for v4 need their v5 versions, which also bring their translations.
The full list of breaking changes and step-by-step upgrade notes are in the v5.0 release notes.
If you are evaluating Aikeedo for the first time, there has never been a better moment. Canvas, templates, the community feed and translations for every plugin are all included out of the box.
Try the live demo to see it in action, or head to our pricing page if you are ready to get started.
More to come.
Discover more on related topics that might interest you