AI in the composer
- Generate: writes a caption for you. Type your idea first, or leave it empty and pick "Post idea" to have the AI propose one.
- Adapt to network: rewrites your existing text in the style and length of a specific network.
- Translate: translates your text.
- Hashtags: suggests up to 5 relevant hashtags.
- Image: generates an image from a text prompt, or edits an image already in the post.
- Video: generates video, if you have added a video provider key in Studio → Your own keys. The Studio also gathers everything the AI creates (video, image, voice and sound), the voices and the styles in one place.
The AI always answers in the language you wrote in. If you write your brief in English, the caption comes out in English.
Autopilot goes further: it plans a whole campaign — up to 16 pieces spread over a week or a month, each with a different angle, written in every account's own voice and carrying a suggested date. Everything lands as drafts for you to review; it never publishes directly, and by default no network is ticked. It has its own article in this section.
Whatever the AI writes or designs follows your own posting style, which Ignix learns from the posts your accounts have already published. See the next section.
If your account speaks as a person or as a character rather than as a company, say so in Knowledge and upload a few reference images: from then on the AI writes in the first person and what it draws looks like you. It has its own article in this section.
Everything the AI creates runs on AI credits, included in every plan: 500 a month on Spark, 2,000 on Blaze, 5,000 on Scale and 15,000 on Inferno. An image costs 20 credits (40 if it carries your photos or your character), an Autopilot piece 35 and a caption 2. Each Sparky reply to your customers uses 1 (or 1 per 3 in nodes that only follow their own instructions), and by DM it also counts toward your DMs. Video generation uses your own provider key, with a starting image (image-to-video), pro/1080p quality, duration and more options depending on the provider.
Every prompt for image or video generation is checked against our content policy before it is generated. If a prompt is rejected, rewrite it and try again.