← Back to open source9 articles·≈ 27 minutes
WeKnora: grounded answers, scoped memory and deliberate operations
Nine chapters from first document ingestion to event-driven RAG, candidate-pool recall, deployment, cost and sandbox boundaries.
Latest article
Follow the documented plugin pipeline without confusing internal stage completion, SSE delivery and source-supported answer quality.
WeKnora RAG architecture: events, shared state, retrieval stages and resumable outputWeKnoratencent/weknora
Start reading →By publication date
01 → 09
01·3 min read·Search intent: Learning
Follow the documented plugin pipeline without confusing internal stage completion, SSE delivery and source-supported answer quality.
WeKnora RAG architecture: events, shared state, retrieval stages and resumable outputWeKnoratencent/weknora
·3 min read·Search intent: Learning
Select the smallest knowledge workflow that answers the task, and choose Lite or team deployment by operating requirements rather than feature count.
Choose WeKnora retrieval, Quick Q&A, Agent or Wiki by the user’s intended outcomeWeKnoratencent/weknora
·3 min read·Search intent: Learning
Separate standard Compose, Lite and development mode; understand cached images, migration risk and why full profiles are optional.
Deploy WeKnora with a known image version, bounded services and a tested data planWeKnoratencent/weknora
·3 min read·Search intent: Learning
Use a small synthetic handbook and an unanswerable question to test the path from upload to source-supported answers.
Your first WeKnora knowledge base: verify parsing, retrieval and citations separatelyWeKnoratencent/weknora
·3 min read·Search intent: Learning
Design a read-only learning project that links source versions, recall behavior and claim-level acceptance without importing private conversations.
Build a WeKnora evidence workbook: missing passages, memory gaps and answer supportWeKnoratencent/weknora
·3 min read·Search intent: Learning
Understand four different knowledge behaviors and why a self-hosted RAG application is not automatically private or correct.
WeKnora explained: document answers, agents, Wiki and personal memoryWeKnoratencent/weknora
·3 min read·Search intent: Learning
Plan measurements for model calls, queues, semantic recall and vector backfill without converting source constants into speed guarantees.
WeKnora cost and performance: separate ingestion, answer work and memory maintenanceWeKnoratencent/weknora
·3 min read·Search intent: Learning
Distinguish knowledge access from personalization and skill selection from execution isolation, network policy and credential exposure.
WeKnora operations: workspace permissions, personal memory and sandbox authorityWeKnoratencent/weknora
·3 min read·Search intent: Learning
Read scoped vector ranking, candidate growth and reciprocal rank fusion through a concrete low-importance memory regression.
WeKnora memory recall: fix the candidate pool before tuning relevance scoresWeKnoratencent/weknora
