agno-agi shipped Agno v3.0.11 on 2026-09-23 (PyPI agno 3.0.11). `Knowledge.search()` now runs a knowledge-level retrieval pipeline that can widen candidates before reranking, with new `MMRReranker` (relevance vs diversity) and pgvector-focused `RecencyReranker` (score × exponential time decay); adds hostname migration for indexed page sources (dry-run by default), typed `PageCommandResult` with truncation/completeness metadata, and machine-readable `cancellation_stage` on Run/Team/Workflow outputs. Per-vector-db `reranker` is deprecated in favor of Knowledge-level rerankers.

Key Takeaways

  • ✓Shipped: v3.0.11 / PyPI 3.0.11
  • ✓Retrieval: knowledge-level pipeline + MMRReranker (#10115)
  • ✓Recency: RecencyReranker with exponential time decay; page-source migrate dry-run by default
  • ✓Observability: cancellation_stage enum + typed PageCommandResult metadata
  • ✓Migration: deprecate vector-db reranker → Knowledge-level; see Knowledge overview docs
🔗

Project Links & Resources

Direct Access
Direct access to official project resources and documentation
🐙
GitHub Release Agno v3.0.11
github.com/agno-agi/agno/releases/tag/v3.0.11
↗
📖
PyPI agno 3.0.11
pypi.org/project/agno/3.0.11/
↗
🐙
PR #10115 knowledge pipeline + MMR
github.com/agno-agi/agno/pull/10115
↗
📖
Agno Knowledge overview
docs.agno.com/knowledge/overview
↗
📖
Agno Knowledge API reference
docs.agno.com/reference/knowledge/knowledge
↗
🐙
agno-agi/agno repository
github.com/agno-agi/agno
↗
🔬

In-Depth Technical Analysis

Core Background & Industry Pain Points Per-vector-db rerankers force rewrites when swapping stores; near-duplicate chunks crowd diverse hits; hostname moves often mean full reindex; cancellation text matching is brittle across PENDING/EXECUTING/INTERRUPTED. ### Architecture Highlights & Internals v3.0.11 runs a knowledge-level retrieval pipeline in Knowledge.search() (#10115), adds MMRReranker and pgvector-oriented RecencyReranker, page-source inspect/migrate (dry-run default), typed PageCommandResult / MCP structured search results, cancellation_stage on run outputs, optional use_previous_response_id for OpenAI Responses, and Y-API provider. Vector-db reranker is deprecated. ### Authoritative Benchmarks & Measured Scores No public Recall/nDCG or agent success rates. Measure duplicate-rate/diversity with MMR on/off, recency skew, and assert on cancellation_stage enums. ### Developer Hands-on Guide pip install agno==3.0.11; attach rerankers on Knowledge per Knowledge overview and reference. Release: GitHub.