Crosslybot feature

Automatic error correction in posts for cross-posting

Turn it on once for the source channel — and typos won't travel to VK, Max and Odnoklassniki. AI fixes only spelling and grammar, while the meaning and style stay yours.

How it works
all targets get the corrected text
once per post
all targets get the corrected text
result length — otherwise the original is published
±20%
result length — otherwise the original is published

Automatic error correction in Crosslybot is an AI proofreader that checks every post before cross-posting. It fixes spelling and grammar mistakes and nothing else: the meaning, style and length of the text stay yours.

The feature is turned on for the source channel and runs once per post — all targets of the project get the corrected text. The AI response is validated, and if it tried to rewrite the text, the original is published. Other AI features — translation, styles and length changes — are on the AI processing page.

How it works

  1. 1

    Open the source settings

    In the project, pick the source channel.

  2. 2

    Turn on error correction

    One checkbox — and it works for all targets of the project.

  3. 3

    Post as usual

    Before sending, AI checks the post text and fixes mistakes. The original in the source is not changed.

What you get

Spelling and grammar only

AI gets a strict instruction: fix mistakes and do not change the meaning, style, structure or length of the text. Names, channel names and numbers stay as they are.

The result is checked

The AI response is validated: length differs by no more than 20%, the original words are kept, and there are no intros like “Here is the corrected text”. If three attempts fail, the original is published.

Once for all targets

Correction runs at the source level once per post, and all targets of the project get the corrected text.

You pay only for changes

An AI request is counted only if mistakes were actually found and fixed.

Works with translation and style

The original is corrected first, then each target applies its own AI settings — translation or style — to the clean text.

Source edits are respected

If you edit the post while it waits to be published, correction restarts on the new text.

FAQ

What exactly does AI fix?

Spelling and grammar mistakes. The meaning, style, structure and length of the text do not change; proper names, channel names and numbers stay as in the original.

Does correction use AI requests?

Yes, it is a separate AI request, but it is counted only when the text actually changed. Translation and style in targets are counted separately — more about translation.

What if the AI rewrites the text its own way?

Such a response fails the check: the length must differ by no more than 20%, and the original words must be kept. After three failed attempts the original text is published.

Does it work for groups and the Webhook API?

No. Correction is not applied to group sources or to posts that come through the Webhook API — only to channels.

Can I turn correction off for a single post?

Correction applies to all posts of the source. To publish a post without it, temporarily turn the feature off in the source settings.

Which plan includes error correction?

It works within your plan's AI limit, which every plan has, including the free one. Compare plans on the pricing page.

Related features

Free plan

Ready to try?

Connect the bot in 2 minutes. Free plan — no cards, no signup.

More about the bot →