Before you start
You need:
- a BriefNext account (see Plans). You can connect before your plan starts. The tools then answer that the account has no active plan, and they unlock when the plan starts;
- Claude, ChatGPT or Claude Code, with your own account;
- the connector URL:
https://app.briefnext.com/mcp
MCP (Model Context Protocol) is the open standard that lets an AI call outside tools. BriefNext is an MCP server, and your AI adds it as a connector.
Claude
- Open Customize, then Connectors. On a Team or Enterprise plan, an owner adds the connector under Organization settings, Connectors, and you click Connect.
- Click Add, then Add custom connector.
- Name it BriefNext, paste the connector URL, and click Continue.
- Click Add. Sign in to BriefNext and press Allow.
In some versions of Claude, Connectors is under Settings instead of Customize. The other steps are the same.
ChatGPT
- Open Plugins, click the plus button, then Add custom MCP server.
- Name it BriefNext, paste the connector URL, and choose OAuth.
- Confirm the warning and create it. Sign in to BriefNext and press Allow.
Does your ChatGPT show Apps instead of Plugins? Open Settings, Apps, Advanced settings, and turn on developer mode. Then click Create app and paste the connector URL.
Claude Code
Run this in your terminal:
claude mcp add --transport http briefnext https://app.briefnext.com/mcp
Then run /mcp in Claude Code, pick briefnext, and sign in.
A client that cannot sign in can use a bearer token instead. Make one on the panel's Connect page, under Advanced.
Signing in
The first time your AI connects, it opens a BriefNext sign-in page.
- Sign in with your email address and password. Or choose "Email me a sign-in link instead".
- BriefNext shows which app asks for access. Press Allow.
Your AI keeps the connection. You sign in again only when you disconnect, or when the app asks you to. You can see and disconnect your apps on the panel's Connect page.
Pick a profile
An account holds one or more profiles. A profile is one person or one brand, with its own profile data, plan, content, research and spend cap.
- One profile: there is nothing to pick. BriefNext uses it.
- More than one profile: the Allow page asks "Which profile?". Pick one, then press Allow.
One connector works with one profile. This keeps two brands apart: your AI never mixes their data in one conversation. Each answer from BriefNext names the profile it used.
To work with a second profile, add BriefNext a second time with another name, for example "BriefNext Acme", and pick the other profile at Allow. In Claude Code, give the second connector its own name:
claude mcp add --transport http briefnext-acme https://app.briefnext.com/mcp
Your first conversation
Ask your AI: "Use BriefNext. Where do I stand, and what should I do next?" Give it your website address if you have one. BriefNext researches the rest and asks only what it cannot find.
You can see and correct everything BriefNext knows on the panel's Profile page. What you say always outranks what research found.
What each tool does
Your AI picks the tools. You do not need to name them. They follow the order of the work: where you stand, where you are going, what gets you there, what you make, what goes out, and whether it worked. Tools marked paid use paid research and count against your spend cap.
Where do you stand?
get_profile- Reads what BriefNext knows about you, with each gap and why it matters.
update_profilepaid- Fills the profile from your website or from facts you give, and corrects any fact. It is paid when it reads a page or turns typed facts into fields.
assess_sitepaid- Reads and scores each page of your site, and reads its colours and typefaces.
check_visibilitypaid- Checks whether ChatGPT, Claude, Gemini, Perplexity and Google name you when a client asks, and who they name instead.
Where are you going?
get_plan- Reads your plan: where you are, where you want to be, and the steps between.
update_plan- Writes the plan: your position, your target in business terms, and how you will know it worked.
What gets you there?
research_peerspaid- Finds who is winning in your corner, what is overdone, and what works.
whats_freshpaid- Finds what moved this week, and what nobody has explained yet.
find_examplespaid- Finds work from people who walked the same path from unknown to known.
research_keywordspaid- Gets real search volumes for your city, province and country.
update_keywords- Records your choices on keywords: relevant or not, the intent, chosen or dropped.
get_topics- Groups your topics, shows how much demand your content covers, and names the next piece and why.
What exactly do you make?
research_topicpaid- Makes a research brief for one topic: what the top results cover, what they miss, and the page pattern that works.
find_internal_linkspaid- Finds which of your own pages to link to, and where.
render_imagepaid- Makes an image in your brand colours, with alt text.
render_diagram- Draws a labelled diagram: a table, phases, a comparison, a mapping or a timeline.
save_content- Saves a piece your AI wrote, linked to the plan. A page must have an answer at the top, sections and alt text before it is ready to publish.
What goes out, and when?
get_brief- The start of each session: what is waiting, what is due, and one next action.
find_content- Finds what you saved.
get_content- Opens one saved item in full.
delete_content- Deletes a saved item.
export_content- Makes a package to publish: Markdown, HTML, meta data, alt text, images and a copy-and-paste sheet.
Is it working?
get_progress- Compares this period with the last one, and checks you against your success criteria and your plan.
record_result- Records a result you report: an enquiry, a client, a value.
remember- Keeps a decision, a preference or a reaction for next time.
The panel's Activity page shows each tool call, what it did and what it cost.
Spend caps
Some tools buy data or research each time they run. Each profile has a spend cap: a limit on what its paid research can cost in 30 days.
- When a profile reaches its cap, paid research pauses. Your AI tells you.
- Saved data, cached results and free tools keep working.
- The pause ends as older spend leaves the 30-day window.
- The panel's Usage page shows your spend against the cap.
Research results are cached for one day to two weeks, so the same question again does not cost again.
Plans
The tools work when your account has an active plan. Without one, they answer "This account has no active plan." See the plans and how to get access at briefnext.com/#access.
Something does not work? See Support.