Cambrian MCP Integration - Implementation Summary (Updated v2)

Overview

Successfully integrated two Cambrian MCP servers as separate action providers for comprehensive blockchain data and AI-powered research capabilities into the multi-agent trading system.

Architecture Change (November 23, 2025 v2)

Previous Implementation: Single unified provider handling both servers New Implementation: Two separate providers for better separation of concerns

Provider Split Benefits:

Clear Separation: Data operations vs AI/social intelligence
Independent Configuration: Each provider can be configured separately
Selective Access: Agents can use one or both providers as needed
Better Maintainability: Easier to update and debug individual providers
Type Safety: More specific action signatures per provider

MCP Servers Configured

1. Cambrian Data MCP Server

URL: https://cambrian-mcp-server-prod-981646676182.us-central1.run.app/mcp
Purpose: DeFi and on-chain data (Solana/EVM focus)
Capabilities:
- Trending tokens analysis
- Token security metrics
- Price and volume data
- Pool liquidity metrics
- Trader leaderboards
- DEX analytics

2. Deep42 MCP Server (NEW)

URL: https://mcp.deep42.org/mcp
Purpose: AI-powered research and social intelligence
Capabilities:
- AI Orchestration Agent - Multi-tool blockchain research
- Social data and sentiment analysis
- GitHub developer intelligence
- Influencer credibility rankings
- Alpha signal detection
- Project discovery
- Content generation

Implementation Date

Initial: November 23, 2025
Deep42 Integration: November 23, 2025
Provider Split: November 23, 2025 (v2)

What Was Implemented

1. Package Installation ✅

Installed @modelcontextprotocol/sdk@1.22.0 via Bun
Package provides MCP client capabilities for connecting to external MCP servers

2. Separate Provider Architecture ✅

Files:

app/providers/cambrianDataActionProvider.ts (450+ lines)
app/providers/deep42ActionProvider.ts (780+ lines)

Architecture:

Two independent providers for specialized functionality
Each provider connects to its respective MCP server
Unified JSON-RPC interface for all MCP tool calls
Proper error handling and logging for both providers
Agents can use one or both providers as needed

3. Cambrian Data MCP Actions (4 actions) ✅

Provider: cambrianDataActionProvider Class: CambrianDataActionProvider Server: Cambrian Data MCP

getTrendingTokens - Solana trending tokens by price/volume
- Parameters: orderBy, limit, offset
- Returns: Trending token list with metrics
getTokenSecurity - Token security analysis
- Parameters: tokenAddress
- Returns: Security scores, risk indicators
getTokenPriceVolume - Price and volume data
- Parameters: tokenAddress, timeframe (1h-24h)
- Returns: Current price, volume, changes
getTraderLeaderboard - Top traders for tokens
- Parameters: tokenAddress, interval
- Returns: Top traders, statistics, volumes

4. Deep42 AI-Powered Actions (7 actions) ✅

Provider: deep42ActionProvider Class: Deep42ActionProvider Server: Deep42 MCP

Deep42 Actions:

askDeep42Agent ⭐ MOST POWERFUL - AI orchestration agent
- Parameters: question, continueChatId (optional)
- Uses AI to orchestrate multiple tools automatically
- Provides comprehensive blockchain research
- Supports multi-turn conversations
querySocialData - Social data and sentiment agent
- Parameters: question, conversationContext (optional)
- Returns: Alpha signals, sentiment, influencer insights
getInfluencerCredibility - Influencer rankings
- Parameters: minTweets, limit, tokenFocus, timeWindow
- Returns: Credibility scores, track records
detectAlphaTweets - High-alpha tweet detection
- Parameters: limit, tokenFilter
- Returns: Early signals, opportunities, insights
searchProjects - Blockchain project discovery
- Parameters: query, technology, category, chain, socialActivity, limit, hasGithub
- Returns: DeFi, NFT, gaming, infra projects
analyzeTokenSocial - Deep social analysis for tokens
- Parameters: tokenSymbol, daysBack, granularity
- Returns: Sentiment, engagement, influencer mentions
getTrendingMomentum ⭐ NEW - Trending momentum detection
- Parameters: lookbackHours, minVelocity, limit, granularity
- Returns: Tokens with momentum scores, velocities, quality metrics, sentiment
- Business Value: Early trend detection, risk assessment, market timing

5. Provider Registration ✅

File: app/api/agent/prepare-agentkit.ts

Changes:

Imported cambrianDataActionProvider and deep42ActionProvider
Initialized provider instances separately
Added to provider map with distinct keys:
- cambrianData: cambrianDataProvider
- deep42: deep42Provider
Included in default providers list for agents without specific config

6. Cambrian Agent Definition ✅

File: data/agents.json

Updated Agent:

{
  "id": "cambrian-bot",
  "name": "Cambrian Social Researcher",
  "description": "Social data analyst providing insights from Deep42/Cambrian API",
  "strategiesFile": "strategies_cambrian.json",
  "triggerHistoryFile": "trigger_history_cambrian.json",
  "providers": ["cambrianData", "deep42", "pyth", "dexScreener", "walletAssets"],
  "systemPrompt": "...",
  "tone": "analytical",
  "isActive": true,
  "createdAt": 1732400000000
}

Agent Capabilities:

Solana token data (trending, security, price/volume, traders)
Social sentiment analysis
Community engagement metrics
Trending token identification with momentum detection
Project reputation assessment
Influencer credibility tracking
Alpha signal detection
Analytical tone focused on actionable intelligence

7. Data Files Created ✅

Files:

data/strategies_cambrian.json - Empty array (ready for strategies)
data/trigger_history_cambrian.json - Empty array (ready for events)

8. Environment Configuration ✅

File: .env.local

Added:

# Cambrian API for social data via MCP
CAMBRIAN_API_KEY=qvhzvNP8K5sGpmlC0xul4MIKCOGqGRRH

9. Documentation Updates ✅

File: AGENTS.md

Added Sections:

Available Agents - Updated Cambrian Social Researcher (Agent #5)
- Now uses both cambrianData and deep42 providers
- Updated capabilities list
Provider System - Split provider descriptions
- cambrianData: DeFi/on-chain data (Solana/EVM)
- deep42: AI-powered social intelligence
Model Context Protocol (MCP) Integration - Updated section
- Two separate MCP server URLs
- Two separate provider implementations
- Benefits of separated architecture
- Updated available MCP actions (now 11 total: 4 data + 7 social/AI)

Technical Architecture

Separated Provider Pattern

Cambrian Data MCP Server              Deep42 MCP Server
         ↓                                     ↓
HTTP Transport with API Key           HTTP Transport with API Key
         ↓                                     ↓
CambrianDataActionProvider            Deep42ActionProvider
         ↓                                     ↓
AgentKit ActionProvider Interface     AgentKit ActionProvider Interface
         ↓                                     ↓
         └──────────────┬──────────────────────┘
                        ↓
              AI Agents (can call both sets of actions)

Integration Benefits

Standardized Protocol: MCP provides consistent way for AI to interact with external APIs
Separation of Concerns: Two providers, each focused on specific domain (data vs intelligence)
API Logic Isolation: API logic stays on respective MCP servers
Dynamic Capabilities: Tools can be updated server-side without client changes
Flexible Integration: Easy to add to existing agent workflows, use one or both providers
Independent Scaling: Each provider can be updated/maintained independently

Files Modified

New Files:

app/providers/cambrianDataActionProvider.ts (450 lines)
app/providers/deep42ActionProvider.ts (780 lines)
data/strategies_cambrian.json
data/trigger_history_cambrian.json

Modified Files:

app/api/agent/prepare-agentkit.ts (updated imports and provider map)
data/agents.json (updated Cambrian agent providers)
AGENTS.md (updated provider system and MCP sections)
CAMBRIAN_MCP_IMPLEMENTATION.md (this file, v2 update)
.env.local (CAMBRIAN_API_KEY)
package.json (added @modelcontextprotocol/sdk dependency)
bun.lock (updated)

Deleted Files:

app/providers/cambrianActionProvider.ts (replaced by two separate providers)

Usage Examples

Using Cambrian Bot with Both Providers

// User: "What Solana tokens are trending?"
// Agent: Calls getTrendingTokens from cambrianData provider
// Returns: Trending tokens with price changes and volumes

// User: "Show me social momentum for these tokens"
// Agent: Calls getTrendingMomentum from deep42 provider
// Returns: Tokens with high social velocity and quality scores

// User: "What's the sentiment for SOL?"
// Agent: Calls analyzeTokenSocial from deep42 provider
// Returns: Sentiment analysis, engagement metrics, community insights

Adding Providers to Existing Agents

To give any agent access to DeFi data or social intelligence:

{
  "id": "warren-debuffett",
  "providers": ["fearAndGreed", "pyth", "swap", "walletAssets", "dexScreener", "cambrianData", "deep42"]
}

Testing Recommendations

Test Cambrian Data MCP actions (cambrianData provider):
- Query trending Solana tokens: getTrendingTokens({ orderBy: "price_change_percentage", limit: 10 })
- Check token security: getTokenSecurity({ tokenAddress: "So11111111111111111111111111111111111111112" })
- Get price/volume: getTokenPriceVolume({ tokenAddress: "So11111111111111111111111111111111111112", timeframe: "24h" })
- Get trader leaderboard: getTraderLeaderboard({ tokenAddress: "...", interval: "24 HOUR" })
Test Deep42 AI-powered actions (deep42 provider):
- Ask AI agent: askDeep42Agent({ question: "What are the top trending tokens on Solana?" })
- Query social data: querySocialData({ question: "Show me alpha signals for SOL" })
- Detect alpha tweets: detectAlphaTweets({ limit: 5 })
- Get influencer rankings: getInfluencerCredibility({ limit: 10, tokenFocus: "SOL" })
- Search projects: searchProjects({ category: "DeFi", chain: "solana", limit: 10 })
- Analyze token social: analyzeTokenSocial({ tokenSymbol: "SOL", daysBack: 7 })
- Get trending momentum: getTrendingMomentum({ lookbackHours: 24, minVelocity: 1.5, limit: 15 })
Integration testing:
- Start dev server: bun run dev
- Select Cambrian bot in UI
- Test both data-focused and AI-focused questions
- Test trending momentum specifically
- Verify API calls in console logs
- Confirm both providers work independently and together

Known Capabilities

Cambrian Data MCP Tools (50+ tools available)

Solana: trending_tokens, token_security, price_volume, traders_leaderboard, pool transactions, wallet balances
EVM: Aerodrome, AlienBase, Clones, PancakeSwap, SushiSwap, Uniswap V3 pools
Price Data: current prices, hourly/daily aggregations, OHLCV data
DEX Analytics: pool info, liquidity maps, fee metrics, volume analysis

Deep42 MCP Tools (16 tools available)

executeDeep42Agent: AI orchestration with multi-tool access
querySocialDataAgent: Social sentiment and alpha signals
queryGithubAgent: Developer intelligence (direct/enhanced/hybrid modes)
getInfluencerCredibility: Influencer rankings and track records
detectAlphaTweets: High-alpha tweet detection
getTrendingMomentum ⭐ NEW: Trending momentum with velocity metrics
analyzeTwitterUserAlphaMetrics: User-specific alpha analysis
searchDiscoveredProjects: Project discovery across chains
getSocialAssociations: Token-influencer associations
queryRepositoryMarketData: GitHub repo market correlations
getSentimentShifts: Major sentiment shift detection
getTokenSocialAnalysis: Deep social analysis for tokens
executeContentGenerationPipeline: Content generation
executeResearchPipeline (deprecated): Multi-stage research

Key Insights

Why Two MCP Servers?

Specialization & Scalability:

Data Server: Optimized for high-frequency on-chain queries, DEX analytics, price feeds
Deep42 Server: AI-powered analysis, social intelligence, GitHub research

Benefits:

Separation of concerns (data vs AI)
Independent scaling and optimization
Specialized infrastructure for different workloads
Better fault tolerance (one server down doesn't affect the other)

Most Powerful Features

askDeep42Agent (Deep42 provider) - The AI orchestrator that can:
- Automatically select and combine multiple tools
- Provide comprehensive research reports
- Support multi-turn conversations with context
- Adapt to complex, open-ended questions
getTrendingMomentum (Deep42 provider) ⭐ NEW - Momentum detection:
- Early trend detection with velocity metrics
- Quality scores and sentiment analysis
- Risk assessment via engagement patterns
- Market timing insights
- Filter by lookback period and minimum velocity
Social Intelligence Stack (Deep42 provider) - Unique competitive advantage:
- Real-time alpha tweet detection
- Influencer credibility tracking
- Sentiment shift detection
- Project discovery with social metrics
- Trending momentum analysis
On-Chain + Social Fusion - Combine both providers:
- Get trending tokens from Cambrian Data MCP
- Analyze social momentum from Deep42 MCP
- Cross-reference trader leaderboards with influencer credibility
- Make informed decisions with complete picture

Future Enhancements

Enhanced Provider Coordination:
- Cross-provider query optimization
- Caching layer for frequently accessed data
- Unified error handling across providers
Additional Social Metrics:
- Time-series momentum tracking
- Network analysis and influence graphs
- Sentiment trends over time with visualization
Strategy Integration:
- Create social momentum-based trading strategies
- Combine Fear & Greed Index with trending momentum
- Alert system for viral tokens and velocity spikes
- Automated trading based on momentum thresholds

References

Success Criteria - ALL MET ✅

Status

IMPLEMENTATION COMPLETE v2 ✅

All planned features have been implemented successfully, including:

Provider split into two separate providers
Trending momentum endpoint added
Documentation updated to reflect new architecture

The Cambrian MCP integration is ready for testing and use with improved separation of concerns.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Cambrian MCP Integration - Implementation Summary (Updated v2)

Overview

Architecture Change (November 23, 2025 v2)

Provider Split Benefits:

MCP Servers Configured

1. Cambrian Data MCP Server

2. Deep42 MCP Server (NEW)

Implementation Date

What Was Implemented

1. Package Installation ✅

2. Separate Provider Architecture ✅

3. Cambrian Data MCP Actions (4 actions) ✅

4. Deep42 AI-Powered Actions (7 actions) ✅

5. Provider Registration ✅

6. Cambrian Agent Definition ✅

7. Data Files Created ✅

8. Environment Configuration ✅

9. Documentation Updates ✅

Technical Architecture

Separated Provider Pattern

Integration Benefits

Files Modified

Usage Examples

Using Cambrian Bot with Both Providers

Adding Providers to Existing Agents

Testing Recommendations

Known Capabilities

Cambrian Data MCP Tools (50+ tools available)

Deep42 MCP Tools (16 tools available)

Key Insights

Why Two MCP Servers?

Most Powerful Features

Future Enhancements

References

Success Criteria - ALL MET ✅

Status

FilesExpand file tree

CAMBRIAN_MCP_IMPLEMENTATION.md

Latest commit

History

CAMBRIAN_MCP_IMPLEMENTATION.md

File metadata and controls

Cambrian MCP Integration - Implementation Summary (Updated v2)

Overview

Architecture Change (November 23, 2025 v2)

Provider Split Benefits:

MCP Servers Configured

1. Cambrian Data MCP Server

2. Deep42 MCP Server (NEW)

Implementation Date

What Was Implemented

1. Package Installation ✅

2. Separate Provider Architecture ✅

3. Cambrian Data MCP Actions (4 actions) ✅

4. Deep42 AI-Powered Actions (7 actions) ✅

5. Provider Registration ✅

6. Cambrian Agent Definition ✅

7. Data Files Created ✅

8. Environment Configuration ✅

9. Documentation Updates ✅

Technical Architecture

Separated Provider Pattern

Integration Benefits

Files Modified

Usage Examples

Using Cambrian Bot with Both Providers

Adding Providers to Existing Agents

Testing Recommendations

Known Capabilities

Cambrian Data MCP Tools (50+ tools available)

Deep42 MCP Tools (16 tools available)

Key Insights

Why Two MCP Servers?

Most Powerful Features

Future Enhancements

References

Success Criteria - ALL MET ✅

Status