Gumloop

What it does

Gumloop lets teams build AI agents that take a goal, use connected apps, and carry out the work. Agents draw on connectors such as Slack, HubSpot, Salesforce, Gmail and Gong, on written instructions, and on reusable skills. Outputs include drafted emails, CRM updates, research summaries, and documents, spreadsheets or dashboards the agent builds as artifacts. Agents can also open a real browser to sign in, fill forms and download files on sites without an API.

People usually describe a task to an agent, watch it run, then ask it to save its approach as instructions or a skill. Agents can start from triggers (app events, webhooks, schedules) and are reachable in Slack, Microsoft Teams, email, text messages, the web, an API and MCP. What sets it apart is the control layer around agents: per-tool approval settings, spend caps, audit logs, usage insights, and a model router across many AI models.

It suits operations, sales, support and finance teams who want non-engineers to build agents, plus admins who need governance. Pro has a 14-day free trial; Enterprise is by contacting sales. Usage consumes credits, and Pro includes 20k per month. Credit costs vary by model and task, and some agent work still needs human approval or takeover.

What it can do

• Build agents with instructions, connectors, skills and knowledge sources
• Run agents from app, webhook, scheduled or one-time triggers
• Use a real browser with a password vault, human handoff and session recordings
• Delegate work to subagents and generate artifacts such as documents and dashboards
• Set per-tool approvals, spend caps, audit logs and AI model restrictions
• Route tasks across many AI models and track credit spend per agent

Works with

• Slack
• Microsoft Teams
• HubSpot
• Gong
• Gmail
• Salesforce
• Google Drive
• Notion
• GitHub
• Confluence

Example workflows

Churn-risk review

Input: a question about accounts renewing soon. Steps: the agent pulls renewing accounts, compares usage with baselines, scans support tickets and call notes, and drafts follow-ups for approval. Output: a ranked list of at-risk accounts with suggested actions and drafted messages.

Recruiting sourcing

Input: an emailed request to find candidates with a given background. Steps: the agent searches for matches, adds them to the applicant tracking system, and drafts personalized outreach. Output: candidate records and email drafts saved in the requester's inbox.

Before you choose

• Agent work consumes credits that vary by model and task; bringing your own API key can waive model credits
• Sensitive tool calls can pause for approval, so some tasks need a person to respond
• Browser-based tasks may require a human to take over for logins or checks

Listed under

• Coding
• MCP and agent tools
• AI agents
• Enterprise
• Marketing automation
• Mcp
• No-code
• Sales ops

Get started

Enterprise requires contacting sales

1. Start the 14-day Pro trial from the pricing page
2. Create an agent and add the two or three connectors it needs
3. Run a real task, then ask the agent to save its approach as instructions or a skill

Pro includes 20k credits per month plus an 8% orchestration fee; Enterprise has custom credits and discounts. Billing interval for the $37 start price was not confirmed.

14-day free trial.

Source: gumloop.com/pricing, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Gumloop: Build AI agents for work ↗www.gumloop.com
• Pricing | Gumloop ↗www.gumloop.com
• Agents - Gumloop ↗docs.gumloop.com

Company website

gumloop.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.