How to Connect Brika 💚 to Claude and ChatGPT Using MCP

What MCP actually does

MCP (Model Context Protocol) is a standard that lets AI assistants call external tools in real time. When you connect Brika via MCP, your AI doesn't need you to paste any data. It fetches it directly, right inside the conversation.

The practical result: instead of manually pulling stats and summarizing them for Claude, you just ask a question. Claude calls Brika, gets your live data, and answers.

"Which creator in my watchlist grew the fastest this week?" Claude calls get_creator_stats, gets real-time data from your Brika account, and responds with the exact numbers. That whole loop takes under two seconds.

How the connection works

Live data flow once connected

You ask a question

in Claude or ChatGPT

→

AI calls Brika tool

e.g. get_radar_feed

→

Brika returns live data

from your watchlist

→

AI answers with real numbers

no copy-paste needed

Without MCP

Log into Brika, find the stat, copy it, switch to Claude, paste it, ask your question. Repeat for every data point.

With MCP

Ask Claude directly. It pulls live Brika data in the same message. One step, no context switching, always current.

What you need before starting

MCP is available on all Brika plans as an Early Access feature. You need at least one creator tracked in your watchlist so there's actual data to query.

Pick your client below. Claude.ai and ChatGPT use OAuth (handled automatically via a login redirect). Claude Desktop and Cursor require a static API key you generate from your Brika settings.

Four supported clients

Claude.ai

OAuth

Settings → Connectors → Add custom connector
Paste https://www.brika.ai/api/mcp
Click Connect, approve on Brika redirect

ChatGPT

OAuth

Settings → Apps & Connectors → Advanced (Developer mode)
Create connector, paste MCP URL, select OAuth
Approve access on Brika redirect

Claude Desktop

API Key

Generate key at brika.ai/settings → AI Integrations
Add config to claude_desktop_config.json
Quit and relaunch Claude Desktop

Cursor / Windsurf

API Key

Generate key at brika.ai/settings → AI Integrations
Add to .cursor/mcp.json or Settings → MCP Servers
Reload the editor

Step-by-step: Claude.ai and ChatGPT (OAuth)

OAuth is the easiest path. No key to generate, no config file to edit. Brika handles authentication through a login redirect, so the only thing you paste is the MCP server URL.

Setup steps (OAuth clients)

Open Connectors in your AI client

Claude.ai: Settings → Connectors → Add custom connector. ChatGPT: Settings → Apps & Connectors → Advanced (enable Developer mode) → Create connector.

Paste the Brika MCP server URL

This is the same URL for every Brika account:

https://www.brika.ai/api/mcp

Click Connect and approve access

You'll be redirected to Brika. If you're already logged in, it's a single click to approve. You're immediately redirected back to your AI client.

Test it with a quick question

Start a new conversation and type: "Show me my Brika watchlist." If the connection is live, Claude will call get_watchlist and return your tracked creators with their latest stats.

Step-by-step: Claude Desktop and Cursor (API key)

Desktop clients need a static API key instead of OAuth. Generate one from your Brika settings and add it to a config file.

Setup steps (API key clients)

Generate an API key in Brika settings

Go to brika.ai/settings → AI Integrations → Claude & ChatGPT. Name your key (e.g. "Claude Desktop") and click Generate. Copy it immediately since it's shown only once.

Add Brika to your config file

Open your config file and add the following block, replacing YOUR_API_KEY_HERE with the key you just copied:

        {

          "mcpServers": {

            "brika": {

              "type": "http",

              "url": "https://www.brika.ai/api/mcp",

              "headers": {

                "Authorization": "Bearer YOUR_API_KEY_HERE"

              }

            }

          }

        }

Save the file and restart the app

Claude Desktop config: macOS at ~/Library/Application Support/Claude/claude_desktop_config.json, Windows at %APPDATA%\Claude\claude_desktop_config.json. For Cursor, the file is .cursor/mcp.json in your home directory. A full quit-and-relaunch is required, a reload alone won't pick up the new config.

Verify the connection

In Claude Desktop, look for the hammer icon in the toolbar when you start a new conversation. That means Brika tools are live. Ask "What are my top creators by follower count?" to confirm data is flowing.

What you can ask once connected

Brika exposes eight tools to your AI. Each one maps to a real capability inside your account, called automatically when you ask a relevant question.

Available Brika tools

get_watchlist

Lists all your tracked creators with latest follower counts and engagement rates.

get_creator_stats

Deep dive on one creator: growth history, engagement over time, post frequency.

get_radar_feed

Returns recent viral, breakout, and outperforming posts from your watchlist this week.

get_reel_breakdown

AI breakdown of a specific post: hook type, format, timing, and a steal-this tip.

get_content_plan

Your AI-generated weekly content plan: hooks, formats, and best posting windows.

get_best_posting_times

Optimal days and hours to post based on engagement patterns across your watchlist.

add_creator

Adds a creator to your Brika watchlist mid-conversation without leaving the chat.

Keyword search across creators and posts in your watchlist.

Best for growth analysis

Combine get_creator_stats and get_watchlist for side-by-side growth comparisons across all creators you track.

Best for content strategy

Combine get_radar_feed and get_content_plan to connect what's working in your niche directly to your weekly posting schedule.

Three starter questions to try first

Once connected, these are the highest-value queries to run immediately. Each one demonstrates a different dimension of what your AI can now do with Brika data.

"Show me my watchlist" gets the full lay of the land. Creator names, follower counts, engagement rates, latest snapshot. Your AI now knows who you're tracking.
"What's gone viral in my niche this week?" pulls your radar feed and returns the breakout posts with hook patterns and format breakdowns. The fastest way to find steal-worthy content.
"What should I post this week and when?" combines your content plan and best posting times into one actionable answer. No tab-switching required.

What changes when your AI has live data

Context switches saved

per research session

Tools available to Claude

live Brika data tools

Setup time

<5 min

OAuth path is fastest

Quick troubleshooting

If Claude says it doesn't have access to Brika tools: check Settings → Connectors (Claude.ai) or confirm the config file was saved and the app was fully restarted (Desktop/Cursor). A reload is not enough, you need a full quit and relaunch.

No content plan found: generate your first plan at brika.ai/content-system before asking Claude about it.
"Not tracking that creator": the creator needs to be in your Brika watchlist first. Add them at brika.ai/watchlist or ask Claude to run add_creator mid-conversation.
API key stopped working: go to Settings → AI Integrations, revoke the old key, generate a new one, and update your config file.
OAuth redirect didn't return: make sure you're logged into Brika in the same browser. Restart the connector setup from Settings → Connectors.

Connect your AI to Brika 💚 Add creators to your watchlist and let Claude pull live growth data, radar posts, and content plans without ever leaving the conversation.

Try Brika free

How to Connect Brika 💚 to Claude and ChatGPT Using MCP

What MCP actually does

What you need before starting

Step-by-step: Claude.ai and ChatGPT (OAuth)

Step-by-step: Claude Desktop and Cursor (API key)

What you can ask once connected

Three starter questions to try first

Quick troubleshooting

Track creator growth automatically, no spreadsheets required.

More from the blog