About springbrand
SpringBrand connects AI agents to data capabilities and workflows for marketing and go-to-market tasks. It supports research, discovery, monitoring and content generation through a shared entry point, with recurring tasks and unified pay-per-call billing described in the README.
Key features
- 01Monitor social platforms and track trends
- 02Analyze website traffic and SEO
- 03Discover companies, contacts and creators
- 04Generate copy, images, video and voiceover
- 05Combine capabilities into workflows
- 06Schedule recurring monitoring and reports
Use cases
- 01Analyze a competitor's traffic and search keywords
- 02Monitor social discussions about a brand
- 03Find companies and contacts matching a customer profile
- 04Identify creators and plan collaborations
MCP support
Provides a hosted MCP server at https://connector.springbrand.ai/mcp, with a development endpoint also documented. The repository packages Skills and MCP entries for supported hosts; other agents can use the documented Skill-plus-MCP fallback.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://connector.springbrand.ai/mcpclaude mcp add --transport http springbrand-lab-springbrand https://connector.springbrand.ai/mcp{"mcpServers":{"springbrand-lab-springbrand":{"type":"http","url":"https://connector.springbrand.ai/mcp"}}}Open in Cursor{"servers":{"springbrand-lab-springbrand":{"type":"http","url":"https://connector.springbrand.ai/mcp"}}}Open in VS Code{"mcpServers":{"springbrand-lab-springbrand":{"type":"http","url":"https://connector.springbrand.ai/mcp"}}}{"mcpServers":{"springbrand-lab-springbrand":{"serverUrl":"https://connector.springbrand.ai/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMESpringBrand
The AI agent capability marketplace — everything your AI agent needs to take action, all in one place.
English | 简体中文
SpringBrand connects professional data sources, tools, and ready-made workflows to the AI assistant you already use — so it doesn't just answer questions, it executes tasks and delivers results.
The problem it solves
Completing one real task with an AI agent usually means finding tools, registering accounts, configuring APIs, and stitching workflows together by hand. SpringBrand puts those capabilities behind one entry: describe the outcome you want, and your agent chooses and calls the right tools to get it done.
- One entry — no switching between tools, accounts, and platforms.
- Direct delivery — the agent executes tasks and delivers results, not just advice.
- Pay per call — unified billing across tools; no stack of long-term subscriptions.
- Recurring tasks — schedule monitoring, research, and reporting to run on repeat.
Examples
- Given a competitor's URL, analyze its traffic sources, acquisition pages, and search keywords.
- Monitor social media for real discussions about your product, brand, or competitors.
- Find companies and key contacts matching your ideal customer profile.
- Find fitting KOL/KOC and draft actionable collaboration and promotion plans.
- Combine capabilities to produce copy, images, video, and voiceover.
Covered capabilities
SpringBrand covers social platforms including X, TikTok, Instagram, YouTube, Reddit, Pinterest, and 小红书, and can perform: social monitoring, trend tracking, content analysis, audience insight, competitor research, website traffic analysis, traffic source analysis, keyword research, top-page analysis, SEO analysis, prospecting, company and contact discovery, creator discovery, promotion planning, and copy, image, video, and voiceover generation.
These capabilities usually live in separate paid tools. SpringBrand provides capabilities similar to:
- Traffic & SEO research: Similarweb, Semrush, Ahrefs, DataForSEO
- Search & web research: Exa, Tavily, Perplexity, Firecrawl
- Company & contact search: Apollo, People Data Labs
- Creator discovery: WaveInflu
- Content generation: GPT Image, Seedream, Seedance, Nano Banana, ElevenLabs
- Third-party systems: GitHub and other connectors
These product names only illustrate capability scope; they do not imply direct integrations or partnerships.
Capabilities are organized into three domains — Platform, Action API, and Connector. When unsure which applies, ask $ask-springbrand; it recommends exactly one Domain Skill and stops.
Install in one prompt
Paste this prompt into your AI Agent (Claude Code, Codex, Cursor, Copilot, Devin, Windsurf, WorkBuddy, or any compatible Agent); it reads the guide and performs the installation:
Install or update SpringBrand Production by following https://plugin.springbrand.ai/INSTALL.md. Identify this Agent, select one Host client, run the pinned CLI command once, and follow only its structured
nextAction. A key supplied in the prompt stays on the API-key path; with no key, let the Host-native OAuth flow run. After the returned browser or reload action, run the onedoctorcontinuation and report success only when it returnsready.
Or add the hosted MCP server directly:
npx add-mcp 'https://connector.springbrand.ai/mcp'
Development (testing only)
Follow the official SpringBrand development installation guide to complete setup: INSTALL.dev.md Identify the target environment and Host first. Run the pinned CLI command once with the development key if the prompt contains one; otherwise let the Host-native OAuth flow run. Read
nextAction, complete that exact action, run the onedoctorcontinuation, preserve unrelated configuration, and report success only when the result isready.
| Production | Development | |
|---|---|---|
| Install guide | INSTALL.md | INSTALL.dev.md |
| MCP entry name | springbrand | springbrand-dev |
| MCP URL | https://connector.springbrand.ai/mcp | https://devconnector.springbrand.ai/mcp |
| Authentication | API key when supplied; otherwise host-native OAuth | API key when supplied; otherwise host-native OAuth |
| Purpose | Everyday use | Testing only |
Install one coherent release rather than mixing Skill versions. Development is for testing only and should not be used as a production configuration.
When only one client should change, use its single-client guide instead of the all-client entry:
| Client | Production | Development |
|---|---|---|
| Codex | INSTALL.cli.codex.md | INSTALL.cli.dev.codex.md |
| Claude Code | INSTALL.cli.claude-code.md | INSTALL.cli.dev.claude-code.md |
| Cursor | INSTALL.cli.cursor.md | INSTALL.cli.dev.cursor.md |
| OpenCode | INSTALL.cli.opencode.md | INSTALL.cli.dev.opencode.md |
| WorkBuddy | INSTALL.cli.workbuddy.md | INSTALL.cli.dev.workbuddy.md |
Client support
Use the dedicated guide when one exists. Otherwise use the CLI client id with
sp connect <id>, or the Other Agent workflow in INSTALL.md when the host
has no CLI client.
| Host or surface | CLI client | Dedicated guide | Notes |
|---|---|---|---|
| Codex CLI, ChatGPT Desktop (Codex view) | codex | INSTALL.cli.codex.md | Plugin fallback uses the Codex manual guide |
| Claude Code, Claude Desktop Code tab | claude-code | INSTALL.cli.claude-code.md | Plugin fallback: INSTALL.claude.md |
| Cursor | cursor | INSTALL.cli.cursor.md | — |
| OpenCode | opencode | INSTALL.cli.opencode.md | — |
| WorkBuddy | workbuddy-ai | INSTALL.cli.workbuddy.md | Native R2 package |
| Pi | pi | — | CLI client only |
| Devin Desktop or CLI | devin | — | CLI client only |
| Grok | grok | — | CLI client only |
| Hermes | hermes | — | CLI client only |
| omp | omp | — | CLI client only |
| OpenClaw | openclaw | — | CLI client only; Skills install through ClawHub |
| command-code | command-code | — | CLI client only |
| ChatGPT Desktop (Work view), Gemini CLI, DSH, any other host | generic | INSTALL.md Other Agent workflow | Configure the remote MCP entry manually |
Development variants of the dedicated guides are kept in this repository; they are not published to the R2 documentation origin. If a host is not listed, report it as unsupported instead of guessing a configuration path.
What this repository contains
The SpringBrand installer: five Canonical Skills and one MCP entry per environment, packaged as native Plugins for Codex, Claude Code, Cursor, and WorkBuddy. Other AI Agents connect through the documented Skill-plus-MCP fallback.
INSTALL.md and INSTALL.dev.md are written to be read and executed by an Agent. For maintainers: VERSION is the source of truth for releases; release-matrix.json records each channel's pinned Plugin and CLI versions, and python3 scripts/release_matrix.py --check keeps the channel page in sync. Synchronize Skill versions with python3 scripts/sync_skill_versions.py [--check] and validate packages with python3 tests/validate_plugin.py.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.