diff --git a/README.md b/README.md
index dfdb97a..90bffd2 100644
--- a/README.md
+++ b/README.md
@@ -29,3 +29,6 @@ Taking notes on the specialized language models that actually have companies pub
 | **[ShieldGemma 2](https://huggingface.co/google/shieldgemma-2-4b-it) (Mar. 2025)** | [Google/Vertex AI](https://cloud.google.com/vertex-ai) / [Hugging Face](https://huggingface.co/) | • Vertex docs: [Models page](https://cloud.google.com/vertex-ai/generative-ai/docs/models), [ShieldGemma 2 docs](https://huggingface.co/docs/transformers/model_doc/shieldgemma2)<br>• Release: [Gemma releases (Feb 19, 2025)](https://ai.google.dev/gemma/docs/releases) | • [Google AI dev docs](https://ai.google.dev/gemma/docs/shieldgemma) | Safety classifier used by Google themselves in their AI stack |
 | **[Llama Guard 4](https://huggingface.co/meta-llama/Llama-Guard-4-12B) (Apr. 2025)** | [Databricks](https://www.databricks.com/) / [Groq](https://groq.com/) | • Databricks Marketplace: [Llama Guard Model](https://marketplace.databricks.com/details/a4bc6c21-0888-40e1-805e-f4c99dca41e4/Databricks_Llama-Guard-Model)<br>• Groq Cloud catalog: [meta-llama/llama-guard-4-12b](https://console.groq.com/docs/models) | • [HF blog "Welcoming Llama Guard 4"](https://huggingface.co/blog/llama-guard-4)<br>• [Model card](https://huggingface.co/meta-llama/Llama-Guard-4-12B) | Safety guardrails used by major cloud/AI platforms |
 | **[Qwen3-Coder](https://huggingface.co/collections/Qwen/qwen3-coder-687fc861e53c939e52d10) (Jul.–Sep. 2025)** | [Google Cloud (Vertex AI)](https://cloud.google.com/vertex-ai) / [Tinfoil](https://tinfoil.sh/) | • Vertex AI: [Qwen3 Coder in Model Garden](https://cloud.google.com/vertex-ai/generative-ai/docs/release-notes), [Qwen3 Coder docs](https://cloud.google.com/vertex-ai/generative-ai/docs/maas/qwen/qwen3-coder)<br>• Tinfoil: [launch blog](https://tinfoil.sh/blog/2025-09-02-qwen3-coder-private), [changelog](https://docs.tinfoil.sh/resources/changelog) | • [Qwen announcement](https://qwenlm.github.io/blog/qwen3-coder/) | Coding model with enterprise API integration by Tinfoil |
+| **[Schematron](https://schematron.com/) (ISO/IEC 19757-3:2025)** | [NEMSIS](https://nemsis.org/) / [Adobe Experience Manager Guides](https://experienceleague.adobe.com/en/docs/experience-manager-guides/using/user-guide/components/structure-validate/schematron) | • NEMSIS: [National Schematron library + validator](https://nemsis.org/technical-resources/version-3/version-3-schematron/)<br>• Adobe AEM Guides: [Schematron validation workflow](https://experienceleague.adobe.com/en/docs/experience-manager-guides/using/user-guide/components/structure-validate/schematron) | • [Schematron overview](https://schematron.com/document/2755.html)<br>• [Balisage update on ISO Schematron](https://www.balisage.net/Proceedings/vol29/html/Sales01/BalisageVol29-Sales01.html) | Rule-based XML validation standard powering compliance-heavy healthcare and documentation pipelines |
+| **[SWE-grep](https://cognition.ai/blog/swe-grep) (Oct. 2025)** | [Cognition (Devin)](https://cognition.ai/blog/windsurf) / [Windsurf](https://cognition.ai/blog/swe-grep) | • Cognition: [Fast Context subagent deployment](https://cognition.ai/blog/swe-grep)<br>• Windsurf: [Agentic IDE now powered by SWE-grep](https://cognition.ai/blog/windsurf) | • [Introducing SWE-grep & SWE-grep-mini](https://cognition.ai/blog/swe-grep)<br>• [Hacker News discussion](https://news.ycombinator.com/item?id=45607822) | RL-trained code retrieval agent delivering order-of-magnitude faster context gathering for coding assistants |
+| **[SWE-grep-mini](https://cognition.ai/blog/swe-grep) (Oct. 2025)** | [Cognition (Devin)](https://cognition.ai/blog/windsurf) / [Windsurf Fast Context Playground](https://playground.cognition.ai/) | • Cognition: [Mini variant serving 2,800 tok/s Fast Context](https://cognition.ai/blog/swe-grep)<br>• Windsurf playground: [Hands-on deployment](https://playground.cognition.ai/) | • [Introducing SWE-grep & SWE-grep-mini](https://cognition.ai/blog/swe-grep)<br>• [RL for Fast Multi-Turn Context Retrieval](https://news.ycombinator.com/item?id=45607822) | Distilled fast-retrieval model optimized for parallel search with enterprise-grade latency |