Skip to main content

​
Overview

Agent Composer offers pre-built templates so you can get started quickly. Availability is by template type:
  • Basic Search and Agentic Search — available to self-serve users.
  • All other templates in the Template Catalog (e.g. Device Log Analysis, Deep Research, domain-specific templates) are available only for Enterprise. There are no other Enterprise-only features—only access to these additional templates. Request a demo to use them.
Basic Search (also referred to as Fast Search in the template catalog) is a streamlined, lightweight template for quick document retrieval and response generation.
  • Simple architecture: Direct search-to-response pipeline without multi-turn research
  • Fast performance: Minimal configuration for rapid queries
  • Basic RAG: Standard retrieval with citation and groundedness tracking
Use cases: Quick lookups, simple Q&A, straightforward document search. For more templates in this style (e.g. Fast Search YAML), see the Template Catalog.
Agentic Search is a multi-step research template. The agent uses iterative retrieval and tool use to answer complex questions with citations.
  • Multi-turn research: Plans searches, refines queries, and synthesizes answers from multiple retrievals
  • Structured agent loop: Uses defined tools (e.g. search over datastores) with guardrails
  • Citations: Grounded responses with source references
Use cases: Technical documentation Q&A, research over internal docs, customer support with deep lookup. Try the demos: All of these agents use the Agentic Search template (or the same multi-step research pattern). Launch the live demo or read the example page. Guides: Template Catalog for related templates.

​
Device Log Analysis (Enterprise)

The Device Log Analysis template is available in Agent Composer for Enterprise customers. It automates root cause analysis (RCA) on device failure logs (e.g. 3GPP LTE eNB/UE logs) by parsing logs, building searchable databases, and running multi-agent analysis to produce timelines, visuals, and detailed RCA reports. Example page · Launch demo In production: You upload a log file and optionally a debug reference guide, then ask a query (e.g. “Why did the call drop?”). The rest is automatic: the system parses logs, builds databases, runs root cause analysis, can generate Python scripts that run in a secure VM, and produces visuals and a detailed RCA report. This is a longer-running, multi-agent workflow aimed at complex, routine knowledge work. Next steps: Request a demo to get access. Docs and demo: The example page has the full description, example outputs (causation timelines, system health visuals), and report structure. The live demo includes three preloaded sample datasets: 3GPP Wireless, ATE Test Equipment, and Mazda Infotainment (each: log + debug reference). Try it at demo.contextual.ai/log-analysis.