Save your token cost by using shared state
@flowiseai
-
Token Consumption and Model Selection for AI Report Generation
By
–
A few things we’ve learned: – It takes an average of 7 minutes to generate a complete report.
– Consumes a lot of tokens ~1M input, 100K output
– Yes, it’s not suitable for all use cases
– Model selection and fallback options are crucial due to the large amount of findings that -
LLM Agent Determines Research Necessity and Task Planning
By
–
⇄ Condition Agent With the generated report, we let the LLM determine whether additional research is needed or if the findings are sufficient. If more is needed, the Planner Agent reviews all messages, identifies areas for improvement, generates follow-up research tasks, and
-
Writer Agent: Turning Raw Findings Into Structured Reports
By
–
Writer Agent Act as a research writer that turn raw findings into a clear, structured Markdown report. Preserve all context and citations. We find Gemini to be the best for this, thanks to its large context window that allows it to synthesize all the findings effectively.
-
Subagent AI Tools: Web Search, Scraping, ArXiv Access
By
–
Subagent has access to web search, web scrape, and arxiv tools. Google Search – for web search links Web Scraper – for web content extraction. This will scrape the content of the links from Google Search. ArXiv – search, download and read content of arxiv papers
-
Subagents: Spawning Focused Research Tasks with Parallel Capabilities
By
–
🖧 Subagents For each task in the tasklist, a new subagent will be spawned to conduct focused research. Each subagent has:
– Clear task understanding capabilities
– Efficient research planning (2-5 tool calls per task)
– Source quality evaluation
– Parallel tool usage for -
Research Lead Agent: Analyzing Queries and Planning Tasks
By
–
Planner Agent Act as an expert research lead to: – Analyze and break down user queries
– Create detailed research plans
– Generate specific tasks for subagents Output an array of research tasks. Full prompt: https://
github.com/anthropics/ant
hropic-cookbook/blob/main/patterns/agents/prompts/research_lead_agent.md
… -
Multi-Agent AI Research System: Planner, Subagents, Writer
By
–
Overview Planner Agent – analyzes the research query and generates a list of specialized research tasks 🖧 Subagents – creates multiple research subagents, conduct focused research using web search, web scrape, and arxiv tools Writer Agent – synthesizes all findings into
-
Building Multi-Agent Research Systems: Anthropic Architecture
By
–
Last Friday, Anthropic released a gem – How to build multi-agent research system. Over the weekends, our team had tried to re-create the same architecture, not exactly 1-1, but close to it. Here's how it works and what we've learnt:
-
All Integrations Remain Open Source by Default
By
–
Of course, every integrations are always open source