The Creator Workflow, End to End

Feedpik's top tier is sold on turning reading into output, and the individual pieces have their own pages. This is how they connect, because the sequence is where the value is — each stage is worth having on its own and considerably more worth having in order.

The shape

Radar finds it → Tracking follows it → reading collects it → Drafts assembles it.

Four stages, and each one hands the next something it could not have produced alone.

Stage 1: the radar finds the story

Trending Radar surfaces topics several of your sources converged on within 48 hours.

For a writer this is the most useful thing in the product, and not because it finds news first. It finds stories that already have more than one account of them. A scoop nobody else has covered is a link you can share; a story four outlets reached independently is one you can write about, because there is disagreement to work with and enough material to be fair to.

Breadth is 40% of the radar score for exactly this reason. What ranks highest is what most of your sources agreed was worth covering.

The synthesis paragraph on each topic is a fast read on whether it is yours to write. It generates once a day and costs nothing to reopen.

Stage 2: tracking follows it

Once a story is worth your attention for more than a day, a term watch turns it from something you check into something that reports to you.

Set the term with all its aliases — the company, the ticker, the product, the person. A watch that knows one spelling misses most of the story. Case is preserved, so write tickers in capitals.

Two things it gives a writer that the radar does not:

Persistence. The radar is a 48-hour window; a watch accumulates over weeks. Long-running stories are the ones worth writing and the ones a window keeps forgetting.

Sentiment direction. Seven days of how your sources have written about the subject. Read it as the shape of the coverage rather than a verdict on the thing — but "coverage turned in week three" is itself the story, often enough.

If the subject is a person or publication rather than a topic, use a rival watchlist instead: output, cadence and cross-source pickup. Pickup is the one that matters here — it tells you which of their pieces the rest of your feed actually responded to, which is a decent proxy for what landed.

Stage 3: reading collects it

As pieces arrive, Add to draft on each one you might use. Collecting material covers the habit; the essential point is that it happens while reading, not while writing.

Get to at least two sources before you generate anything — one material makes the generator a paraphrase engine, and it inherits the original's framing along with its facts.

Put the ones you want cited in the first eight. Everything past that is attached and listed but does not reach the generator.

Stage 4: Drafts assembles it

Generate draft produces a working title, a hook, three to six sections and a takeaway, with every factual claim carrying [1]-style markers back to the material it came from. 8 credits, charged only if a draft actually comes back.

Then Copy with sources, which is the step people skip and should not. Hand-selecting the text carries the [n] markers and leaves behind the list explaining them — a piece full of orphan citations with the attribution lost in transit.

What the draft is and is not

It is a first draft, assembled from sources you chose, with its citations intact. It is a structure to argue with and a guarantee that every claim in it traces to something you read.

It is not your piece. The generator is told to write only from the materials, which is what keeps it honest and also what caps it: it cannot have the opinion, notice what is missing, or know which of two accounts is more credible. Those are the parts you were going to do anyway, and the ones worth your time.

What it costs

Per piece, roughly: radar free, tracking free to run, 8 credits for a generate, a few more if you regenerate. Max includes 4,000 credits a month, which is deliberately sized for someone doing this several times a week rather than occasionally.

If something looks wrong

The radar never has anything to write about — likely too few sources, or sources spread across too many subjects. Two of them must cover the same story within 48 hours. Concentration beats breadth for this.

My watch collects nothing — check spelling and case, then check that sources you follow would plausibly cover the subject. A watch reads your feed; it does not search the web.

The draft reads like the first source — you probably had one material, or the second was past the eighth position. Check what actually reached the generator.

Regenerate wiped my edits — it replaces the editor's contents and there is no undo. Copy before regenerating.

I have Pro, not Max — Radar is yours; Tracking and Drafts are not. The first two stages work on Pro and produce a reading list rather than a draft. See plans.