About LexSocket¶

What is LexSocket?¶

LexSocket is an AI-ready platform providing unified access to European public procurement data. We aggregate tender information from TED (EU above-threshold) and 11 national procurement platforms (below-threshold), making them searchable through a single MCP (Model Context Protocol) server.

Our mission is to democratize procurement intelligence — enabling companies, researchers, and consultants to discover opportunities and analyze public spending across Europe using AI agents.

The Problem We Solve¶

European public procurement is fragmented:

• TED (Tenders Electronic Daily): EU-wide above-threshold tenders (€140k+ for services, €5.4M+ for works)
• 11 National Platforms: Below-threshold tenders published only on country-specific sites

Finding opportunities across this landscape requires: - Manually querying multiple platforms in different languages - Parsing inconsistent data formats - Combining results manually - Monitoring for new tenders continuously

LexSocket solves this by: - Aggregating 11 countries into a single search interface - Supporting hybrid (full-text + semantic) search - Providing advanced filtering (CPV codes, NUTS regions, value ranges, deadlines) - Keeping data fresh with automatic daily updates - Enabling AI agents to automate the entire workflow

Data Coverage¶

TED (Above-Threshold)¶

• Source: Official Journal of the European Union (TED)
• Coverage: All EU/EEA countries (31 countries)
• Data Types: Procurement notices, contracts, planning notices
• Update Frequency: Twice daily (06:00 UTC, 18:00 UTC)
• Latency: 2-4 hours from publication
• Historical Data: Available back to 2014

National Tenders (Below-Threshold)¶

Aggregated from 11 national platforms:

Technology Stack¶

Data Pipeline¶

Source APIs → Collector → Parser/Normalizer → SQLite (source of truth)
                                                    ↓
                                             BGE-M3 Embeddings
                                                    ↓
                                              LanceDB (search index)
                                                    ↓
                                            Google Cloud Storage
                                                    ↓
                                            MCP Server (Cloud Run)

Key Technologies¶

• Indexing: SQLite (source of truth) + LanceDB (vector + FTS search)
• Search:
• Full-text search (BM25 via Tantivy)
• Vector search (BGE-M3 embeddings, 1024 dimensions)
• Hybrid (Reciprocal Rank Fusion)
• Embeddings: BGE-M3 (BAAI/bge-m3) with ONNX INT8 inference
• Infrastructure: Google Cloud Platform (Cloud Run, Cloud Storage)
• Authentication: Auth0 OAuth2 + API keys
• API Standard: Model Context Protocol (MCP)

Key Features¶

Hybrid Search¶

Combine full-text keyword matching with semantic vector search for better results. LexSocket automatically ranks results using Reciprocal Rank Fusion.

Rich Filtering¶

• By CPV code: Common Procurement Vocabulary (product/service classification)
• By NUTS region: Geographic location of work
• By contract value: Min/max estimated value
• By deadline: Find tenders closing soon
• By buyer: Search by contracting authority name
• By date: Publication date or submission deadline ranges
• By procurement type: Open, restricted, negotiated procedures
• By status: Active, awarded, cancelled (National Tenders only)

Similar Tender Discovery¶

Found one opportunity? Get similar tenders using semantic similarity. Perfect for market research and opportunity mapping.

Procurement Statistics¶

Analyze trends by country, sector (CPV), region (NUTS), procurement type, or currency. Great for market research and competitive analysis.

Use Cases¶

For Businesses & Consultants¶

• Opportunity hunting: Find government contracts in your sector
• Market research: Analyze procurement spending by country, region, sector
• Competitive analysis: Monitor competitor wins and market trends
• Bid management: Track upcoming deadlines and similar opportunities
• Sales enablement: Identify strategic accounts with procurement activity

For Researchers & Journalists¶

• Public spending analysis: Investigate government procurement patterns
• Transparency: Track where public money goes
• Sector analysis: Study trends in infrastructure, IT, healthcare, etc.
• Legislative impact: Analyze effects of policy changes on procurement

For Procurement Officers¶

• Benchmarking: Compare pricing and terms across European markets
• Best practice research: Learn from similar tenders
• Regulatory compliance: Monitor procurement requirements across countries
• Vendor intelligence: Track buyer behavior and preferences

Data Quality & Freshness¶

Update Schedule¶

Data Accuracy¶

• TED: Data comes directly from official TED sources (100% accuracy)
• National Tenders: Data parsed from official national platforms
• France (BOAMP): Custom JSON eForms format
• Germany: OCDS releases with CPV/NUTS extraction
• Spain: Atom XML feeds with detailed metadata
• UK: OCDS releases from Contracts Finder
• Others: Platform-specific formats normalized to common schema

Known Limitations¶

• NL (TenderNed): CPV and NUTS codes not available in summary endpoint; estimated values often rounded
• Data lag: National platforms may update at different times; expect 12-24 hour lag
• Status updates: Status (active/awarded/cancelled) reflects last known state, may lag actual changes
• Multilingual: TED supports multiple EU languages; National Tenders typically in national language

Security & Privacy¶

Data Protection¶

• HTTPS/TLS: All connections encrypted in transit
• Authentication: OAuth2 via Auth0 or API keys; no passwords stored
• API keys: Cryptographically secure, can be rotated instantly
• PII: Tender data is public procurement information; no personal data collected or stored beyond what's published

Compliance¶

• GDPR: Compliant with EU data protection regulations
• Data location: Servers in EU region (Google Cloud Belgium)
• Retention: Data retained for 12 months; older data archived

Roadmap & Future Features¶

Planned Enhancements¶

• More countries: Expanding to cover additional European nations
• Real-time alerts: Get notified when matching tenders publish
• AI-powered summaries: Automatic tender summaries and bid analysis
• Historical analysis: Trend analysis and forecasting
• Price intelligence: Estimate contract values and benchmarks
• API rate tier options: Higher limits for heavy users
• Batch export: Download large result sets

Pricing¶

LexSocket offers a free tier with standard API rate limits (60 requests/minute, 10,000/day). Contact us for enterprise tiers with higher limits and support.

The Team¶

LexSocket is built by a team passionate about EU procurement transparency and AI. We combine expertise in:

• European public procurement law and practice
• AI/ML and natural language processing
• Cloud infrastructure and data pipelines
• Open data and regulatory compliance

Open Source & Community¶

• GitHub: github.com/lexsocket
• Public Data: All data is aggregated from public sources
• MCP Standard: Built on the open Model Context Protocol standard
• Contributing: We welcome feedback, bug reports, and feature requests

Contact & Support¶

Get Help¶

• Email: support@lexsocket.ai
• Documentation: lexsocket.ai/doc/
• GitHub Issues: github.com/lexsocket/issues
• Twitter: @lexsocket

Report Issues¶

Found a bug or have a feature request? Please report it on GitHub Issues or email us.

Feedback & Suggestions¶

We'd love to hear how you're using LexSocket. Share your feedback at support@lexsocket.ai.

License¶

LexSocket is licensed under the MIT License. See GitHub repository for details.

FAQ¶

Is the data real-time?¶

No, but it's fresh. TED updates twice daily (6+ hours latency), National Tenders daily (same-day to next-day). For real-time monitoring, subscribe to official platform notifications.

Can I download all tenders?¶

Not through the API, but you can export results. Contact us for bulk export options.

What languages are supported?¶

TED tenders are published in multiple EU languages. National Tenders are typically in the national language. Search works across all languages using semantic search.

Can I use this for commercial purposes?¶

Yes, the data is public. You can use LexSocket data for commercial research, bid analysis, and market intelligence.

How long does data persist?¶

At least 12 months of historical data is available through the API. Archived data is available on request.

Next Steps¶

• Get Started: Connect in under 5 minutes
• Search Guide: Master procurement search
• API Reference: Detailed parameter and error documentation