Niche research

Niche research turns a few sentences about what you want into a list of channel ideas, each judged honestly for whether it can work as a faceless channel that ChannelForge animates from code. Use it when you want to choose the niche yourself — for one channel, or for a whole network of channels from one brief.

The one-click setup on the start page runs the same research and takes its best pick for you. This page is about doing it by hand, in a project, where you see every proposal and decide.

What it does#

A project is a brief and the channels built from it. When you create one, a research job reads the brief and proposes niches — ten on the first run — each with:

Each proposal is weighed on six questions: is there demand for it on Shorts and Reels today, how crowded is it and what angle is free, can code render it well (most attractive niches fail here — "celebrity facts" cannot be drawn, "how a lock works" can), is there depth for 300 episodes or more, what is the risk (medical or legal advice, scare content, politics, religion, anything aimed at children, anything that needs a licence, anything that reads as misinformation, or a channel that looks like a brand's own), and is it entertaining — built on what most viewers do not know, with consequences shown on familiar things at human scale. The research is told not to invent view counts.

It also picks a shortlist — as many niches as the project wants channels — that work together as a set, and says why under Why the shortlist works together. Shortlisted proposals carry a shortlist badge. Each run picks a shortlist of its own and earlier badges stay, so after Propose more the shortlist can hold more niches than Channels wanted, and the reason shown is the latest run's.

How to use it#

1. Open Projects and press New project (editors, admins and owners).
2. Fill in the form — see Settings below — and press Create and research niches. You land on the project's page; the research usually takes under two minutes, and the page updates by itself.
3. Read the Niche proposals. For more, press Propose more: eight new proposals that do not repeat the earlier ones.
4. Draft the ones you like:
   • Draft this channel on one proposal. In about a minute it becomes a channel in Draft: its name, YouTube, Facebook and Instagram handles, tone of voice, palette, tagline, at least eight places its episodes rotate through, and its first topic ideas. The proposal then shows channel drafted.
   • Draft the N shortlisted drafts every shortlisted niche that has no channel yet, in one click, within your plan's channels.
5. Build them. Under Channels on the project page, Set up the N drafted channels (Set up the drafted channel when there is one) gives each drafted channel its own setup: its kit is built, its first episode made, its brand suite rendered and two more episodes made, side by side as GPUs free up. You get an email as each first episode is ready. Or open a channel and press Build the kit, then Build the brand and Generate, one at a time.

Prefer a renderer that already exists? Add a channel from a format on the project page starts a proven format inside this project, in its language.

Writing a good brief#

The brief steers every proposal. Say what kind of channels, for whom, what is on screen and what to avoid:

Faceless science channels for curious adults: physics and engineering made visual, every
episode built on a surprising fact shown at human scale. No politics, no medical advice.

• Be specific. "Ancient engineering: how the Romans built roads, aqueducts and domes, shown being built" gives sharper niches than "history".
• Name what you will not do. Disaster footage, health claims, children's content — the research takes it out.
• Think in pictures. Everything is drawn from code: machines, space, oceans, cities, bodies, maps and simulations render well; real people, celebrities and news events do not.
• Ask for a network. Set Channels wanted to 4 and the shortlist is four channels that are different from each other, not four takes on one idea.

Settings#

The New project form:

On the project's page, Pause project stops all of its channels from generating and posting until Resume project.

Credits and plans#

Research, drafting and building are ordinary jobs, each with its price beside its button: what it typically costs and what it reserves. Niche research and a channel draft cost a few credits each; the renderer kit is the large one-time cost of a custom channel. See the What things cost table on Pricing and Credits and plans.

Niche research is on every plan. Your plan sets how many projects and channels the workspace may have; Draft the N shortlisted stops at the plan's channels and says how many did not fit.

Limits and good to know#

• The scores are an informed judgement, not live platform data and not a forecast of views or income. Use them to compare ideas and choose what to try first.
• A drafted channel starts with its schedule off and posting times of 09:00, 15:00 and 21:00 in the workspace's default time zone. Nothing posts until you connect accounts and turn the schedule on.
• Each channel is drafted to be different from the workspace's other channels; the research sees them.
• A setup started from the start page with Or describe your own channel cannot go into a project where another such setup is running: both would choose the project's best niche. The project's own Set up the N drafted channels is different: each drafted channel already has its niche, so they run side by side.

Troubleshooting#

Related#

• Proven formats — skip the research and the kit build
• The free niche finder — three ideas from one sentence, with no account
• Channels and kits — what happens after the draft
• Outlier radar — the competitors a niche's research names, watched