Gumloop

What it is

Gumloop is a no-code AI workflow canvas where ops teams drag, drop, and connect modular nodes to build multi-step AI automations without writing code. A flow is a visual graph: a trigger node fires on a schedule or an event, data moves through enrichment nodes, AI model nodes (Claude Sonnet, GPT-4o, Gemini 2.5, DeepSeek — each selectable per node), and action nodes that write back to the apps you use. The platform ships 130+ native integrations, 115+ pre-made blocks, and a Chrome extension that uploads browser-side data directly into flows. The separation from standard workflow automation (Zapier, Make) is document intelligence — structured extraction from PDFs, emails, and images at scale — and per-node model flexibility: a single flow can call Claude for legal clause extraction, GPT for copy generation, and Gemini for image classification, with model cost optimized per step.

Gumloop launched in YC W24 (Winter 2024), raised $70.1M across a $3.1M seed (First Round, July 2024), a $17M Series A (Nexus Venture Partners, January 2025), and a $50M Series B led by Benchmark (March 2026). Named customers include Shopify, Ramp, Gusto, Samsara, and Instacart. The company has 37 employees and is headquartered in San Francisco.

Why it shows up in RevOps, Legal, and Recruiting stacks

The same canvas covers all three verticals because the underlying job is identical: take unstructured inputs, run AI over them, push structured outputs into the tools that need them.

• RevOps uses Gumloop for prospect-enrichment pipelines: pull a list from HubSpot, scrape each company website, run a Claude node to extract ICP signals, and write enriched fields back to the CRM — no code required. It also runs for email outreach personalization (ingest list → LLM-personalize body → export to sending tool) and CRM data-hygiene jobs that would otherwise require a Python script or an ops engineer.
• Legal ops sets up document-processing flows to ingest PDFs from a shared folder, extract clause data or party names with a GPT node, and route flagged documents to a Slack alert or contract intake inbox.
• Recruiting automates candidate screening pipelines: ingest applicant CVs, score against a job description via LLM, push qualified rows into the ATS, and notify the recruiter — replacing what is otherwise a 2-3 hour daily manual sort.

Pricing reality

Gumloop runs two public tiers and a custom enterprise plan:

• Free — $0; 5,000 credits/month, 1 seat, 1 active trigger, 2 concurrent runs. Designed for prototyping a single flow, not for production team use.
• Pro — $37/month; 20,000+ credits, unlimited seats, unlimited active triggers, 5 concurrent runs, 25 concurrent agent interactions, 1 hosted MCP server instance, team collaboration. The production tier for most teams.
• Enterprise — custom; adds role-based access controls, SCIM/SAML SSO, VPC deployment, AI spend insights, workflow queuing, multiple MCP server instances, and a dedicated support channel with an embedded Gumloop expert.

At $37/month for 20,000+ credits versus Lindy’s $49.99/month for 5,000 credits, Gumloop delivers roughly 4× the credit volume at a lower price. The comparison is imperfect: Gumloop credits run a canvas you build yourself; Lindy credits power pre-built judgment agents. Budget against the specific flow you intend to run, not the headline ratio.

Best for

Non-engineer RevOps, Legal Ops, and Recruiting teams in the 5–300-person range that need document-heavy AI pipelines, prospect-enrichment automations, or multi-LLM workflows that no single point-tool covers. ROI is clearest when replacing ≥1 manual research or processing task that currently consumes 3+ hours per week at a per-seat software cost that exceeds $37/month.

Alternatives and when to pick them

• n8n — the leading self-hosted open-source workflow automation platform. Pick it when your team has engineering support, needs full data residency, and wants to own the infrastructure. n8n self-hosting is free; cloud plans start at $24/month. Connector depth exceeds Gumloop’s, but document-intelligence nodes require custom code rather than pre-built blocks.
• Lindy — pick when the automation needs persistent judgment: watching an inbox, triaging inbound, or reasoning across CRM state over time. Lindy’s agent model is always-on and event-driven; Gumloop’s canvas runs flows to completion on a trigger. For lead routing or meeting-note summarization, Lindy’s judgment-first design fits better than a canvas.
• Zapier — the market-share leader in workflow automation, with the widest integration catalog. Pick it when the work is deterministic plumbing (move data from A to B when C fires) and no AI reasoning step is needed. Zapier’s per-task pricing scales poorly for AI-heavy pipelines at volume compared to Gumloop’s $37/month flat rate.
• Make — pick when complex branching logic, a wide connector library, and lower per-operation cost at high volume are the priority over AI-native document intelligence. Make’s free tier is more generous than Gumloop’s; it’s the fastest-growing alternative in the budget-automation segment.

Watch-outs

• Credit burn is front-loaded in AI-heavy flows. LLM calls, document parsing, and image analysis nodes consume credits far faster than simple data-movement operations. A flow processing 200 PDFs through a Claude Opus node can exhaust the Pro plan’s 20,000 credits in a single run. Guard: build the flow, run it on a 10-row test set, read the credit dashboard, then size your plan from the observed per-row cost — not from the headline number.
• The 5-concurrent-run cap on Pro creates a throughput ceiling for large batch jobs. A 500-row list processed in one batch finishes in 100 times the single-flow duration when concurrency is capped at 5. Guard: measure batch completion time against the window your downstream systems require; move to Enterprise (which adds workflow queuing and higher concurrency limits) if the gap is unworkable.
• MCP server hosting is capped at 1 instance on Pro. Teams that need to expose multiple internal APIs via MCP — for Claude Code agents, cross-flow data routing, or AI spend controls — need Enterprise. Guard: if MCP is a requirements driver for your architecture, get Enterprise pricing before building the integration.