AI replies: draft responses to social mentions
Draft on-brand public replies and DMs to mentions with response profiles for goal, tone and examples. You review every draft and post it yourself.
Last updated
AI replies draft an answer to a mention so you never start from a blank box. The reply composer at the bottom of every mention writes a public reply or DM. You edit it, copy it and post it yourself. SnitchFeed never posts for you.
Drafts follow a response profile: a shared preset for your goal, tone and example replies. Set one up first, then generate replies from any mention.
Brand Context required
Drafting needs a workspace Brand Context. If yours is missing, set it up under Workspace Settings → Brand Context first. See Brand Context.
Credits
| Action | Cost |
|---|---|
| Generate a draft | 5 credits |
| Rewrite a draft (Shorten, More human and so on) | Free |
See Credits.
Response profiles
A response profile tells the AI what you want from a thread and how the reply should sound. Profiles are shared with your workspace. Keep separate ones when people or motions need a different voice, such as support and founder-led replies.
Create a profile
Go to Workspace Settings → Response Profiles. You can also click Set Up in the reply composer if your workspace has no profiles yet.
Click Create a new Profile. Name it at the top of the editor and pick a Profile Color.
Fill in Instructions & Goal, Voice Settings and Voice Examples, described below.
Click Save. If you added examples, SnitchFeed learns your voice from them as it saves.
Instructions & Goal
Response goal is required. Pick what the reply should achieve:
| Goal | Use it to |
|---|---|
| Drive conversion | Move toward a next step, such as a trial or demo, when the intent supports it. |
| Build relationship | Write warm, peer-to-peer replies that build trust without a pitch. |
| Ask for more info | Ask clarifying questions before recommending anything. |
| Handle feedback | Acknowledge friction or requests and offer help. |
| Provide support | Solve the problem first, before any sales angle. |
Choose Custom to write your own goal, for example "Offer a technical comparison without naming competitors".
Special instructions is optional free text: persona, topics to stress, phrases to avoid or compliance notes.
Voice Settings
Four sliders set the defaults for every draft made with this profile:
| Slider | Steps |
|---|---|
| Tone | Casual, Relaxed, Balanced, Formal, Very formal |
| Length | Very short, Short, Medium, Long, Very long |
| Assertiveness | Hedged, Cautious, Balanced, Direct, Bold |
| Brand plug | No plug, Subtle name, Name only, Name + link, Soft CTA |
Brand plug controls how much your brand may appear:
| Step | The reply may include |
|---|---|
| No plug | No brand name, link or call to action. |
| Subtle name | Your brand name once, very subtly, with no link. |
| Name only | Your brand name, with no promotional line. |
| Name + link | Your brand name plus an allowed link. |
| Soft CTA | Your brand name plus one soft call to action, such as a trial, demo or doc, using allowed links only. |
When Brand plug allows it, CTA goal sets the next step to steer toward: None, Book demo, Drive website visit, Drive DM, Drive engagement or Custom.
Voice Examples
Optional. Add example replies for Reddit, X, LinkedIn and Bluesky, up to five per platform with Add example. Use short replies you have actually posted. They work better than polished marketing copy.
Without examples, drafts rely on your goal, sliders and Brand Context.
Edit or delete a profile
Select a profile on the Response Profiles page, make changes and click Save. To delete it, use the delete option in the editor and confirm Delete profile?. Drafts already made with it stay on their mentions.
Generate a reply
The composer appears on every open mention, in feeds, Opportunities and collections.
Open a mention and scroll to the composer, or press R.
Pick a profile in the profile picker at the bottom of the composer.
At the top, choose a public reply (Comment or Tweet, depending on the platform) or DM.
Click Generate Comment, Generate Tweet or Generate DM, or press G. The draft appears in the text box.
Edit the text if you like, then click Engage or press E. SnitchFeed copies the reply and opens the post in a new tab so you can paste it.
Generate again for another version. The Draft 1/3 arrows switch between versions for this mention.
Generating or rewriting a draft marks the mention Engaged and Seen for your team. See mention states.
Public reply or DM
| Platform | Public reply label | DM |
|---|---|---|
| Comment | Yes | |
| X (Twitter) | Tweet | Yes |
| Comment | Yes | |
| Bluesky | Tweet | Yes |
| Hacker News | Comment | No |
Public replies and DMs keep separate drafts. Switching tabs brings back the last draft for that mode.
Fine-tune one reply
Open Fine-tune Response to change settings for this mention only. They start from the selected profile.
| Control | Range |
|---|---|
| Goal | A preset such as Drive conversion, or Custom with your own text. |
| Tone | Casual to Very formal. |
| Length | Very short to Very long. |
| Confidence | Hedged to Bold. Called Assertiveness in the profile editor. |
| Brand Plug | No plug to Soft CTA. |
When your settings differ from the profile, buttons next to the profile picker let you keep them. Update profile saves them to the profile after you confirm. Duplicate copies the profile with your changes, and New creates a fresh profile from them. Edit opens the full profile editor.
Rewrite a draft
Buttons under the text box rewrite a draft without starting over. Each rewrite adds a new version and is free.
| Rewrite | Effect |
|---|---|
| Shorten | Tighter copy, same intent. |
| Expand | More detail and context. |
| Rephrase | Same meaning, different words. |
| Less salesy | Tones down promotional language. |
| More human | Warmer, more conversational. |
| More technical | More precise, deeper detail. |
Rewrites turn off when your fine-tune settings no longer match the draft. Generate again, or switch back to a version that matches.
Character limits
The composer counts characters and turns the count red when you go over:
| Platform | Limit |
|---|---|
| Bluesky | 300 |
| 1,000 | |
| 1,250 | |
| X (Twitter) | 28,000 |
| Hacker News | No limit |
Public replies on X are drafted to stay within 300 characters. Run Shorten if a draft runs long.
What the AI reads
Each draft uses the mention and its context (such as the parent post of a Reddit comment), the listener that found it, your Brand Context, and the profile's goal, sliders, CTA goal and examples.
Tips
- One motion per profile. Keep sales and community support in separate profiles.
- Keep Brand plug low on Reddit and X unless the thread clearly invites a product mention.
- Switch to DM when support details should stay private.
- Try a variant safely. Use Duplicate in the composer to test new settings without changing the original.
Next steps
- Opportunities: find the mentions most worth a reply.
- Intent & Brand Context: the company profile every draft uses.
- Feeds: filter to High Fit Score or Buying Intent mentions to reply to.