Sirsh Amarteifio·Oct 6Use Jev for what its best at: Semantic JoinsEntity resolution, knowledge graphs and recommendations all need judgements we can afford to repeat across large collections.
Haydar Külekci·Aug 24, 2025Understanding Elasticsearch Percolator Field and Ingestion LifecycleElasticsearch offers powerful features for full-text search, but one lesser-known and incredibly useful tool is the Percolator. Unlike…
Sirsh Amarteifio·Apr 30, 2025Building an AI memory proxyIn Percolate we use many different patterns for wiring together LLMs, data and clients for different contexts. One way I like to think…
InPercolation LabsbySirsh Amarteifio·Apr 8, 2025OpenWebUI as Percolate ClientToday I hooked up the OpenWebUI client to Percolate so I’m leaving a note about the parameters that I use. This is quick to do once you…
InPercolation LabsbySirsh Amarteifio·Mar 27, 2025Agent Orchestration: The basics with PercolateI wanted to discuss a few things about testing agent orchestration (see also the notes). I want to discuss general ideas as they relate to…
InPercolation LabsbySirsh Amarteifio·Mar 21, 2025Local free embeddings may still be slow but its never too early to start thinking about themWe live in an exciting time of technology but its also strange in many ways. AI is shaking up a lot of the industry and the economy, and…
Sirsh Amarteifio·Mar 21, 2025The Emerging Agentic Operating SystemAt the heart of working with Large Language Models is the message context. This is a relatively small message payload we send to the LLM…A response icon1A response icon1
InPercolation LabsbySirsh Amarteifio·Mar 11, 2025Model Context Protocol & Percolate REST bindingsMCP, open sourced by Anthropic, is an open protocol that allows AI agents to securely interact with local and remote resources through…
InPercolation LabsbySirsh Amarteifio·Feb 2, 2025Setting up K8s on a dedicated Hetzner “Robot” ServerI am documenting my experience setting up for the first time with Hetzner a dedicated server provider that has a great service at great…A response icon1A response icon1
InPercolation LabsbySirsh Amarteifio·Jan 31, 2025Generating LLM response structures in multiple schemes in PercolateI spend a lot of time testing generating “message stacks” for talking to LLM Apis and tool calling. Because OpenAI, Google and Anthropic…A response icon1A response icon1
