Gumloop

Gumloop is built for the automations a solo founder actually dreams about delegating: give it an outcome, not a recipe. Its agent builder lets you describe a worker — research every new lead, summarise what the company does, judge fit against your criteria, draft a personalised opener, and drop the result in your sheet or CRM — then connect the tools it may use and let it run on a schedule, a trigger, or from a chat prompt. Founders use it for lead research and enrichment batches, inbox and support-draft preparation, competitor and market monitoring, content research pipelines, and document-heavy chores like reading PDFs or contracts and filing structured summaries. A friendly in-product assistant helps assemble flows from plain language, and the legacy visual workflow builder still covers deterministic jobs where an agent's judgement is unnecessary. The important difference from Zapier or Make is the cost shape: Gumloop unifies model tokens, compute and paid tool calls into credits priced at a published per-credit rate, plus a small orchestration fee, so an AI-heavy run costs what the underlying model work costs rather than a flat per-step fee. That is honest pricing, but it means two runs of the 'same' agent can cost different amounts depending on how much reading and reasoning the input needed. There is no standing free plan — entry is a 14-day Pro trial — which suits founders who already know the job they want automated.

Billed per credit: credits unify model tokens, compute and paid tool calls at $0.005 per credit, plus an 8% orchestration fee on each run (16% if you bring your own model key). Pro includes 20,000 credits/month; overage is opt-in pay-as-you-go at $0.005/credit.