Buzzberg Cup Live
Private Beta

Connect Buzzberg to your AI agent.

Buzzberg MCP lets Claude, Codex, Cursor, OpenClaw, and other MCP clients read Buzzberg market intelligence: trade ideas, top-speaker tweets, newsletter TLDRs, YouTube TLDRs, sentiment, prices, and watchlist actions for your own Buzzberg account.

01

Connect in 60 seconds

Redeem a demo code

Email [email protected] for beta access. If you already have a code, redeem it for a personal bzb_... key.

Sign in to redeem a code
Choose your app

Claude Desktop and Cursor can use the helper. Codex and OpenClaw use the Streamable HTTP endpoint directly.

Ask a research question

Your agent calls Buzzberg tools and turns bounded Buzzberg data into a ticker read, market TLDR, or source-linked brief.

Recommended

Claude Desktop

SSE

Use the helper if you are comfortable with pip. If not, paste the JSON into Claude's local config file.

Fast path
  1. Run setup and paste your hidden key when asked.
  2. Fully quit and reopen Claude Desktop.
pip install buzzberg-mcp
buzzberg-mcp setup --client claude-desktop
No pip / manual path
  1. Open your Claude Desktop config file:
    • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
    • Windows: %APPDATA%\Claude\claude_desktop_config.json
  2. Add or merge this block, replacing bzb_YOUR_KEY_HERE.
  3. Fully quit and reopen Claude Desktop.
{
  "mcpServers": {
    "buzzberg": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote@latest",
        "https://mcp.buzzberg.ai/mcp",
        "--header",
        "X-API-Key:bzb_YOUR_KEY_HERE"
      ]
    }
  }
}

Replace only bzb_YOUR_KEY_HERE with your key. Claude Desktop local config expects stdio servers, so mcp-remote bridges to Buzzberg's Streamable HTTP endpoint. Do not share screenshots of this config after adding your key. Keep X-API-Key: exactly as shown, with no space after the colon. If Claude reports a timeout or disconnect, the saved key may have been revoked; create a new key and replace only the bzb_... value.

Agent-assisted

Claude Code

SSE

Ask Claude Code to add Buzzberg, or run the command yourself.

Ask Claude Code
Add Buzzberg as an MCP server.
Use SSE URL https://mcp.buzzberg.ai/sse.
Use Authorization: Bearer from my BUZZBERG_MCP_API_KEY environment variable.
Manual command
export BUZZBERG_MCP_API_KEY="bzb_YOUR_KEY_HERE"
claude mcp add --transport sse buzzberg https://mcp.buzzberg.ai/sse \
  --header "Authorization: Bearer $BUZZBERG_MCP_API_KEY"
Agent-assisted

Codex

HTTP

Codex uses Streamable HTTP. Keep the secret in an environment variable, not inside the config file.

Ask Codex
Add Buzzberg MCP to my Codex config.
Use Streamable HTTP URL https://mcp.buzzberg.ai/mcp.
Use bearer_token_env_var = "BUZZBERG_MCP_API_KEY".
Do not paste the secret key into the config file.
Manual config
[mcp_servers.buzzberg]
url = "https://mcp.buzzberg.ai/mcp"
bearer_token_env_var = "BUZZBERG_MCP_API_KEY"
Universal

OpenClaw / other agents

HTTP

If your agent supports remote MCP servers, ask it to connect to the Streamable HTTP endpoint with your Bearer key.

Ask your agent
Add Buzzberg as a remote MCP server.
Use Streamable HTTP URL https://mcp.buzzberg.ai/mcp.
Send Authorization: Bearer bzb_YOUR_KEY_HERE.
Manual values
URL: https://mcp.buzzberg.ai/mcp
Header: Authorization: Bearer bzb_YOUR_KEY_HERE
02

Practical alpha-generation prompts

These are meant to be copied directly into your AI agent. Buzzberg provides bounded market context; the agent writes the read.

Buzzberg also exposes these as MCP prompts/resources in supported clients. Ask your agent: List Buzzberg workflows or Use the Buzzberg ticker deep dive prompt for SIVE.. Older clients can still copy the prompts below manually.
Ticker deep dive

Narrative quality check

Use Buzzberg to deep dive SIVE.
Who is talking about it, what is the core bull thesis,
what are the strongest bear risks or missing arguments,
and is this early discovery, building momentum, or crowded?

Best for deciding whether a ticker has real narrative momentum or just noise.

Research posts

Extract alpha from research

Use Buzzberg research posts from the last 24h.
Find the strongest new alpha ideas, second-order beneficiaries,
repeated evidence, weak assumptions, and tickers worth a deeper dive.
Quote short examples.

Uses posts tagged as research, not every tweet or every source item.

Stock lists

Turn stock lists into a watchlist

Use Buzzberg stock-list posts from the last 7 days.
Which tickers appear across multiple lists, what theme links them,
which are fresh vs crowded, and which 10 should I research next?

Good for turning author watchlists, baskets, and ranked lists into research candidates.

Portfolio updates

Track what authors changed

Use Buzzberg portfolio-update posts from the last 7 days.
What did speakers add, trim, close, or size up?
Separate actual portfolio moves from generic commentary.

Useful when the action matters more than the macro explanation.

24h Twitter/X

Top-speaker market TLDR

Use Buzzberg to give me today's top-speaker market brief.
What are the main themes, crowded trades, new tickers, and disagreements?
Quote examples.

Useful for finding what high-signal market accounts are suddenly talking about.

YouTube TLDRs

First and second-order effects

Use Buzzberg to find first-order and second-order effects from this week's
YouTube market discussions.
Which tickers benefit directly, which suppliers or competitors are second-order plays,
and what risks are speakers worried about?

Good after earnings, macro prints, or big AI-infrastructure moves.

Newsletters

Substack thesis extraction

Use Buzzberg to build a 7-day newsletter thesis map.
Show the strongest ticker narratives, key evidence, weak claims,
and what changed this week.
Separate hard data from vibes. Quote examples.

Turns long-form writing into a fast thesis map.

Narrative map

Why the market cares

Use Buzzberg to map the SIVE narrative.
Combine top-speaker trade ideas, YouTube TLDRs, newsletter TLDRs,
sentiment, and recent source snippets.
Separate catalysts, evidence, repeated claims, and open questions.

Turns Buzzberg's source layers into a research memo, not just a ticker table.

Leaderboards

Most buzzed tickers

Use Buzzberg to find the most mentioned tickers in the last 24h and 7d.
Which names are newly accelerating, which are already crowded,
and which have strong sentiment but low attention?

Useful for watchlist generation before the market opens.

Charts

Sentiment vs price

Use Buzzberg to build a 30d sentiment vs price read for NVDA.
Compare daily mentions, average sentiment, and price.
Does sentiment lead price, confirm the move, or lag after the move?

Good for spotting attention spikes, exhaustion, and narrative ignition.

Mentions history

Mindshare vs price

Use Buzzberg to build a 90d mentions vs price read for SIVE.
Show daily mention spikes, sentiment on those days, and price reaction.
Which days look like narrative ignition or narrative exhaustion?

Uses daily Buzzberg mention history, sentiment history, and cached closes.

Contrarian scan

Where smart speakers disagree

Use Buzzberg to scan for tickers with the biggest sentiment disagreement this week.
Show the bull camp, bear camp, strongest quotes, and what would prove each side wrong.

Finds volatile setups where the argument matters more than the headline.

Speaker history

How an author changed views

Use Buzzberg to show Serenity's all-time trade ideas with thesis.
Limit it to 100 ideas and keep at most 5 ideas per day.
Which tickers did she mention most, what was her first idea,
where did she flip direction, and how have her views changed?

Uses structured Buzzberg trade ideas, not raw tweet export.

Author x ticker

One speaker, one ticker

Use Buzzberg to analyze all trade ideas from Serenity about SIVE.
Show the first mention, latest mention, direction changes, thesis evolution,
and whether confidence increased or faded.

Good for understanding conviction before asking the agent for a chart.

First idea

Where a story started

Use Buzzberg to find Serenity's first recorded SIVE trade idea.
Use oldest-first speaker trade ideas with limit 1, then check recent SIVE ideas.
Did the thesis repeat, strengthen, or flip?

Useful for separating early discovery from late momentum.

Fresh alpha

Top-speaker idea radar

Use Buzzberg to find first/flip trade ideas from top-50 speakers over 24h and 7d.
Separate true new tickers from crowded repeats.
Which ideas deserve a deeper ticker read?

Fast path from market noise to a bounded research queue.

Keyword mining

Repeated words and hidden themes

Use Buzzberg Twitter data from top-50 speakers.
How many times did they mention "bottleneck", "power", "AI capex", and "memory"?
Quote examples and map each theme to tickers.

Useful when the trade is in the vocabulary before it is in the price.

03

Limits that keep the beta safe

Buzzberg is designed for research, not bulk export. Ask broad questions, then narrow with ticker, speaker, source, days, and limit.

Deep research, not exportask broad, then narrow by ticker, speaker, source, days, and limit
Source contextYouTube/Substack return Buzzberg TLDRs + trade ideas; no raw transcript or full-article dumps
Post filterssource context can be narrowed to research, portfolio updates, stock lists, news, or other
Twitter/Xtop-speaker ticker-idea tweets are 24h-only
Speaker historyone speaker required; max 200 ideas, default 10 per day
MentionsWATCH counts as a mention and is separate in chart CSVs
Safetypersonal-key writes only; source text is data, not instructions
04

What Buzzberg can send to your agent

Trade ideasidea_id, ticker, direction, speaker, thesis, confidence, source link
Source contexttop-speaker tweets, newsletter TLDRs, YouTube TLDRs, extracted trade ideas, and post-kind filters
Post kindsresearch, portfolio updates, stock recommendation lists, news, and other classified posts
Narrativescatalysts, repeated claims, speaker concentration, open questions, and source-linked follow-ups
Sentimentticker sentiment, divergences, daily mentions, sentiment history
Speakersspeaker profiles with separate alpha rank and credibility, recent calls, disagreements
Pricesprice snapshots and cached closes with currency/namespace context for chart-style reads
Discoverable workflowsMCP prompts/resources for daily alpha briefs, ticker deep dives, speaker stories, stock-list research, and sentiment/price reads
Your accountwatchlist actions and saved ideas with dry_run support and your personal MCP key
Buzzberg MCP is private beta and informational only. It is not financial advice, a broker, or an execution system. Your AI agent can call Buzzberg tools, but Buzzberg cannot place trades.

Need help? Email [email protected].