Introducing "Scale Mode": your AI agent can now do 100,000s of tasks

Quick Overview

Nelima introduces "Scale Mode," a breakthrough feature enabling the AI agent to execute incredibly long, complex tasks consisting of hundreds or thousands of steps autonomously from a single prompt, overcoming the context window limitations that plague other AI agents.

Key Points: Scale Mode allows Nelima to execute virtually endless chains of actions (hundreds or thousands of steps) autonomously from one single prompt. This capability overcomes the context window limitation that typically breaks other AI agents during long tasks. Nelima maintains coherence by storing task states, intermediate results, and decisions externally, allowing it to remember progress and next steps. A demonstration showed Scale Mode processing a complex prompt: finding 5,000 B2B SaaS companies based on strict criteria, gathering specific data (name, domain, CEO info) for each, scoring them 1-10, and saving results to a database. The co-founders, James Kachamila and Arjon Das, emphasize that this feature enables a 10x output increase for knowledge workers across every domain. Users are encouraged to upgrade to the Pro plan ($40/month) for features like High Usage Limit and Extended Long Running Scale Mode, which supports these token-heavy operations.

Context: James Kachamila and Arjon Das, co-founders of Nelima, introduce a major new capability called "Scale Mode." This mode addresses the fundamental limitation of current AI agents, which often fail or lose context when attempting long, multi-step tasks. Scale Mode allows Nelima to handle massive workloads by maintaining state and execution history externally, enabling complex, autonomous workflows that previously required constant human intervention.

Detailed Analysis

Nelima's new "Scale Mode" represents a significant advancement in AI agent capabilities, allowing it to run virtually endless chains of actions (hundreds or even thousands of steps) based on a single prompt, bypassing the restrictive context window limitations found in other agents. This persistence is achieved because Nelima stores task states, intermediate results, and decisions externally, allowing it to remember what has been done and what needs to happen next without losing coherence, even across massive sequences of operations. Co-founder James Kachamila provided an example where Nelima was instructed to find 5,000 B2B SaaS companies meeting strict criteria (industry, location, size, founding date), then gather specific data points for each (name, domain, CEO info), score them 1-10 based on fit, and finally batch-insert all results into a Supabase database—all within that one initial prompt and executed in Scale Mode. Co-founder Arjon Das contrasted this with other agents that forget context or require human input for long tasks. The presenters noted that due to the high token usage associated with these long-running tasks, users might need to upgrade from the Free plan to the Pro plan ($40/month) to access features like High Usage Limit and Extended Long Running Scale Mode.