The export button is missing on the free plan. Everyone in our group hit it within ten minutes of signing up.
ReproducibleFelo X Search on Grok Bot
Grok Bot
Grok BotA launch watch that reports complaints, not mention counts
A Bot watching a launch overnight can collect hundreds of posts and still tell you nothing in the morning. Felo X Search returns what was posted with the signal each post carries, so the routine reports what people are actually running into.
Save this as a skill called "Launch watch". Group what people post by signal, quote the reproducible reports, and do not count mentions.
- Core units
- Bot · skill · routine
- Query
- Product, handle, or launch
- Comes back as
- Posts with a signal tag
- Cadence
- Launch week, then weekly
A Grok Bot owns a job, a conversation and a schedule, and a skill is the method it runs. Felo skills are the research half of that loop, saved once and reused by every routine.
- Core units
- Bot · skill · routine
- Skills come from
- A saved run, or Settings → Plugins → Yours
- Runs on
- iPhone, Android, and desktop
Posts grouped by what they claim, which is what a morning report needs
The overnight run should hand you four findings, not four hundred posts. Sorting by signal is what makes that possible, and it keeps one loud thread from reading as a trend.
Migration is manual and there is no way to batch it. Support confirmed that is intended, which is not what the announcement implied.
Expectation gapNo complaints from us, but we only use the API, so we skip the whole onboarding path.
Scope limitSharing the launch post again with no added detail.
Echo, not reportingWrite the rule into the saved method: include reproducible reports and expectation gaps, drop the echoes, and never report a count as a finding.
Add it on Grok BotLearn it during launch week, keep it afterwards
The best moment to teach this method is while something is actually happening, because the examples you approve become the rules the routine follows.
- Run
Give one Bot a real job
Create a Launch Watch Bot with a narrow description: it reports what was posted, it never replies, and nothing leaves the account without approval.
- Save
Ask it to record the method
Run one pass during the launch, correct the grouping, then save the method with the classification rules you just applied by hand.
- Enable
Turn the skill on where it is needed
Enable it for the launch Bot, and leave it off any Bot that speaks in public channels, because a classification is easy to mistake for a decision.
- Schedule
Attach a routine, then read the first run
Attach a routine for the week after the launch, then a weekly one. Read the first run and confirm it dropped the echoes instead of listing them.
Three ways a launch watch misleads
The first day of a launch produces more heat than signal, and the method should say so.
A post is a lead, not a finding
One reproducible report outranks a hundred complaints. The signal tag is what keeps that ranking visible in the morning summary.
Volume is not severity
A loud thread about a button can bury a quiet one about billing. Sort by what the post claims, not by how many people repeated it.
Launch week is not normal week
Report daily while it matters, then weekly. A daily routine that continues past the launch is how the reports stop being read.
Two requests for launch week
Both fit a routine that runs overnight and reports into the Bot's thread in the morning.
Pull what people posted about the launch in the last 24 hours. Group by signal, quote the reproducible reports, and ignore the reposts.What comes backA short morning list of things people actually hit, with quotable posts behind each one.
Is support responding publicly? Quote the replies from official accounts with timestamps.What comes backA clear picture of whether the company is answering, or an honest statement that nobody official has replied.
Three checks before the morning report is trusted
The overnight gap makes verification more important, not less: nobody watched the run happen.
Open one post from the report
Find it on X. A handle that does not resolve is the fastest catch for a set that was assembled rather than collected.
Check the dates against the launch
Everything quoted should be from the period being reported on. An older post in a launch report shifts the story.
Ask what was dropped
Request the posts that were classified as echoes. If there are none, the filter is not working and the report is just a feed.
Teach the Bot once, read the launch every morning
Run it by hand during launch week, save the method, then let the routine keep watch while you sleep.
What Grok Bot users ask about Felo X Search
Can the Bot reply to complaints?
The skill only reads. Replying is a separate capability, and keeping them apart is deliberate: a classification is not a decision about how to respond.
How often should the routine run?
Daily during a launch, weekly afterwards. A daily watch that continues indefinitely produces reports nobody reads by the third week.
Does it need my X account?
No. It reads an index through the Felo CLI, so nothing about it touches your own timeline or login.
What about posts in other languages?
Ask for the grouping in your working language and keep the quotations in the original, so a quote can still be checked against the post.
Every Felo skill for Grok Bot
Five skills written for Grok Bot. They follow the same pattern on this platform, so you can combine them in one session without re-explaining your setup.
Felo Search
Felo Slides
Felo YouTube Subtitling
Felo SuperAgentOne API key. A full agent toolkit.
Start with a skill, then compose the exact models, live data, knowledge, and output tools your Grok Bot workflow needs — the same key every Felo skill already uses.
Models
Use cost-efficient LLM access when your agent needs custom reasoning or generation.
- LLM API
- Chat completions
- Model access
Live information
Give agents current web and social signals instead of relying on stale context.
- Web Search
- X Search
- Web Fetch
Knowledge and output
Ground work in reusable knowledge, then turn it into shareable deliverables.
- LiveDoc
- PPT generation
- Mindmaps
Creative tools
Extend the workflow from research to visual and office-ready results.
- Image generation
- Page tools
- Office outputs