We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
Remove Phase 8: Enterprise Features from roadmap
Update roadmap years: 2025→2026, 2026→2027
Add Roadmap link to main Wiki page
Add development roadmap
docs: Clean up Home.md - remove links to non-existent pages - Remove links to pages that don't exist in Wiki - Keep only essential navigation links - Simplify structure for clarity - Update Telegram-Safety-Guide.md: clarify LLM usage
docs: Translate remaining Russian text in API-Reference.md
docs: Fix broken links in Deployment.md
docs: Fix Wiki documentation issues - Replace PATAS-core references with PATAS (Deployment, Quick-Start, Code-Overview) - Fix PATTERN_MINING_CHUNK_SIZE default (1000 -> 10000) in Configuration - Fix broken links in Quick-Start (Telegram-Integration, Troubleshooting) - Translate Russian text to English in API-Reference - Update all repository URLs to kiku-jw/PATAS
docs: Fix LLM comparison in Comparison-and-Positioning - Clarify that PATAS uses LLM for offline pattern discovery, not real-time classification - Fix comparison with Classic LLM Filters: different LLM usage patterns - Add key difference explanation: offline discovery vs real-time classification - Update integration section to clarify PATAS role - Add 'Data Volume' row to comparison table
docs: Simplify LLM Usage page - remove redundant information - Remove repetitive 'offline only' and 'NOT used for real-time' statements - Fix deployment options: clarify on-premise vs external LLM (OpenAI sends data externally) - Remove obvious redundant information - Remove 'What LLMs Do NOT Do' section (obvious) - Remove 'LLM vs Machine Learning' section (not relevant) - Remove 'LLM Limitations & Mitigations' section (covered in validation) - Make content more concise and focused - Keep only essential information
docs: Improve LLM Usage documentation - Remove section emphasizing LLM is not needed (too negative) - Strengthen description of LLM role: deep semantic understanding, significantly improves accuracy - Clarify LLM doesn't just generate SQL, but understands semantics and variations - Add LLM rule validation before production (adequacy and false ban risk assessment) - Add sensitive information encryption before sending to LLM - Emphasize LLM's crucial role in pattern analysis accuracy
docs: Add links to new documentation (scaling, two-stage, benchmarks, production reports)
docs: Add Scaling and Cost Design page New page explaining how PATAS handles large-scale processing: - Chunked processing - Two-stage pipeline - Fast clustering (DBSCAN) - Batching & caching - Aggressiveness profiles - Simple mode fallback - Cost comparisons Tested on 20K real Telegram messages. Shows 50-80% API cost reduction. Also updated Home.md: - Added link to new Scaling page - Fixed repository link (PATAS-core → PATAS)
Initial wiki: Complete PATAS Core documentation