Skip to main content
Agents are AI-powered assistants that use tools to solve open-ended tasks. Unlike workflows that follow a fixed path, an agent decides which tools to use and when, adapting its approach to the task in front of it.

​
What agents do

You give an agent a goal and a set of tools, and it figures out how to get there: which tools to call, in what order, and when to ask you for input.
  • Adaptive: takes different approaches for different situations
  • Tool-driven: uses your apps and workflows as needed
  • Conversational: works through tasks in a back-and-forth chat
  • Context-aware: considers your instructions, skills, and conversation history
When you give it a task, the agent analyzes the request, decides which tools to use and in what order, runs them, adapts based on the results, and asks for confirmation when your instructions call for it.

​
Configure an agent in three moves

You do not need to write a long system prompt up front. The fastest way to a useful agent is to let it do real work, then have it write its own instructions.
1

Connect your apps

Open the Apps section and add the integrations the agent needs (Gmail, Salesforce, Slack, and so on). The apps you connect define what the agent can see and do, so start with the two or three it actually needs.
2

Start chatting and give it a real task

Use the built-in chat to run an actual task end to end. Watch where it guesses, asks the wrong thing, or misses a step.
3

Ask it to write its own instructions and skills

Once it completes a task the way you want, tell it: “Update your system prompt so you always do it this way” or “Turn this into a skill.” The agent writes its own instructions and skills, getting you 90% of the way without hand-authoring anything.
This loop is the whole point: correct the agent once, have it capture the correction in its prompt or a skill, and it stops making that mistake. Every conversation makes the agent better.

​
The agent builder

The configuration panel has two tabs:
  • Agent: everything the agent is made of. Agent Preferences holds the model and the system prompt (its instructions), and below that sit Triggers, Apps, Skills, Subagents, and Abilities.
  • Settings: how the agent presents and behaves outside its core logic: Personalization, Agent Details, Chat Preferences, Slack Preferences, Secrets, and the Danger Zone.
The rest of this page covers each part, starting with the ones you will reach for most.

​
Apps

Apps are the integrations your agent connects to, such as Gmail, Salesforce, Slack, Notion, and 150+ more. The apps you connect determine what the agent can access, so this is the most important part of configuration.
Apps section showing connected app icons, an Add an App row, and the AI Discovery toggle set to ON
Click + App to open the picker. The All tab lists every available app (Gumloop-managed and your own), and the Custom tab filters to custom MCP servers you have added. Pick the few your agent actually needs rather than connecting everything.
Add an app modal with a search box, All and Custom tabs, a Connected list with checkmarks, and an All apps list
The AI Discovery toggle on the Apps header is the same setting as Tool Discovery in Abilities. It lets the agent load tool schemas on demand instead of all at once, which keeps context lean when you connect many apps.

​
Account selection

Click any connected app to open its detail view. The Account selector controls which login the agent uses to call that app’s tools.
Google Docs app detail view showing Activity, Tools Enabled, and Rules tiles, with an Account selector offering Use Personal Default and Use Specific Account
  • Use Personal Default: each person who runs the agent uses their own default account. This is the default.
  • Use Specific Account: pin one account for the agent, useful when you have multiple accounts for the same service.
  • Use Team Default (team agents only): everyone on the team uses the same shared account.

Credentials

How agents authenticate, personal vs. team credentials, and the account selection flow.

​
Tool Management and approvals

The same detail view has a Tool Management panel that controls which of an app’s tools the agent can call, and which ones need your approval first. Tools are grouped into Read-only tools and Write/delete tools.
Tool Management panel with an approval preset dropdown, read-only and write/delete tool groups, and per-tool allow, ask, and deny controls
Set a preset for the whole app, or control each tool individually:
PresetBehavior
Always allowEvery tool runs without asking.
Ask each timeThe agent pauses for approval before any tool call.
Ask for writes/deletesRead-only tools run freely; write, delete, and unknown-risk tools pause for approval.
CustomSet the mode per tool: allow always, ask each time, or never allow (deny).
When a tool needs approval, the agent pauses mid-task and shows you an approval card. This is the human-in-the-loop guardrail for sensitive actions.

Human in the Loop

How approvals work, what the agent shows you, and how to tune when it pauses.

​
Skills

Skills teach your agent how to do specific work your way: multi-step processes, templates, and domain knowledge that load only when relevant. You rarely write one by hand. The easiest path is to prompt the agent in a chat: get it to do a task well, then say “turn this into a skill.” When it gets something wrong later, give it feedback in the chat and it updates the skill. The AI Skill Editing toggle is what lets it create and edit skills on its own.
Skills section with AI Skill Editing toggle and the Skill menu showing Create With AI, Upload Files, Write Skill Instructions, and Add Existing Skill
You can also add a skill yourself with + Skill:
  • Create With AI: describe the skill and the agent generates it.
  • Upload Files: turn a document or .zip into a skill.
  • Write Skill Instructions: enter a name and description yourself.
  • Add Existing Skill: attach a skill you already created.

Agent Skills Guide

Create skills, attach them to agents, and build a library that improves over time.

​
Triggers

Agents can run on their own, without you starting the chat. The Triggers header shows AI Managed: when on, the agent can create, edit, and manage its own triggers and schedules during a conversation.
Triggers section with AI Managed toggle and the Trigger menu showing Create With AI, App Trigger, Scheduled Trigger, and One-Time Trigger
Click + Trigger to add one of four kinds:
OptionWhat it does
App TriggerRuns the agent when an event happens in another app (new email, new Slack message, a record changes).
Scheduled TriggerRuns the agent on a recurring schedule, for example every weekday at 9 AM.
One-Time TriggerRuns the agent once at a specific time.
Create With AIDescribe what you want and the agent builds a custom trigger for you.

Agent Triggers Guide

Set up app and scheduled triggers, write prompt templates, and manage active triggers.

​
Subagents

Subagents let your agent delegate to other agents. Instead of doing everything in one conversation, it can spin up focused helpers that work in parallel, then collect the results and continue.
There are two ways an agent delegates, both through the invoke_agent tool:
The agent clones itself, keeping the same tools and instructions. Useful for parallel work: spawn several clones, each handling a different subtask. The clone shows up as “(Me)” in the list and is enabled by default. Clones cannot clone themselves again (depth limit of 1), and you can scope a clone to a subset of apps.
Each subagent runs as its own conversation with its own context and sandbox, visible in your chat history. How many can run at once depends on your subscription tier. Behind the scenes, subagents run as queued background tasks with their own time budget (about half the parent’s). For batch invocations a shared progress board tracks each one, and the parent can hand specific files to a subagent before it starts. The parent reads each subagent’s results when it finishes.

​
Abilities

Abilities are the agent’s built-in capabilities. Most are on by default, and each can be toggled from the Abilities section.
Abilities section listing Web Search, Web Fetch, Image Generation, Search Past Conversations, Ask Question, Tool Discovery, and App Rules Creation, with a Workflow button
AbilityDefaultWhat it does
Web SearchOnSearches the web for current information. Choose the provider (Exa, Parallel, Firecrawl, or the model’s native search).
Web FetchOnReads the content of a specific URL. Choose the provider (Firecrawl, Parallel, Exa, or Gumloop).
Image GenerationOnCreates images from text prompts.
Search Past ConversationsOnSearches and retrieves earlier conversations for context. The backbone of an agent that learns over time.
Ask QuestionOnLets the agent pause and ask you a structured, multiple-choice question when it needs a decision.
Tool DiscoveryAutoLoads tool schemas on demand. In Auto, the agent loads tools directly when they are small (roughly 10% of context) and switches to on-demand discovery when they grow larger.
App Rules CreationOn where availableLets the agent propose App Rules for your review during a chat. Enabled by default on plans that include App Rules.
Use + Workflow to attach a Gumloop workflow as a tool. The agent decides when to call it, fills in the inputs, and reads the outputs.
When building workflows for agents to call, use clear Input and Output nodes, descriptive names (“Enrich Lead from LinkedIn Profile”, not “Workflow 1”), and keep each one focused on a single job.

​
Code sandbox

Every agent has a built-in code sandbox for running Python and shell commands in a secure, isolated environment. It is always on, so the agent can analyze data, generate files, and run scripts automatically. You do not need to configure anything. The sandbox persists installed packages and workspace files across conversations and ships with 80+ Python packages preinstalled.

Code Sandbox & Secrets

Sandbox capabilities, persistence, execution limits, preinstalled packages, and Agent Secrets.

​
Evaluations and reflections

Two features help you measure and improve agents over time.

Evaluations

Define test cases and grade your agent’s responses so you can catch regressions and compare changes before they ship.

Reflections

Let your agent review its own recent conversations on a schedule and propose improvements to its instructions and skills.
Agents can also produce rich outputs such as documents, spreadsheets, and interactive charts. See Agent Artifacts for how those are generated, shared, and versioned.

​
Embedding agents in workflows

Creating an agent is the 0 to 1. Embedding it in a workflow is the 1 to 100. The Agent node runs any of your configured agents inside a workflow, so you can chain it with other nodes and run it in batch.
CapabilityStandalone AgentAgent in Workflow
Manual chatYesYes
Scheduled and event triggersYesYes
Chain with other nodesNoYes

Agent Node

Embed agents in workflows for chaining and batch processing.

​
Working in chat

​
Chat input menu

The + menu in the chat input bar gives you quick actions while you work:
Chat input plus menu showing Add photos and files, Use skill, Mention integration, Mention secret, and an Incognito toggle
  • Add photos & files: attach files and images to your message.
  • Use skill: manually point the agent at a specific skill.
  • Mention integration: reference a connected app directly in your message.
  • Mention secret: reference a stored secret so the agent can use it without you pasting the value.
  • Incognito: toggle a private conversation (covered below).

​
Voice input

You can send audio instead of typing. Gumloop transcribes it on the server and sends the text to the agent, so the conversation flows the same whether you type or talk.
Agent chat input showing the microphone button for voice input
Supported formats include mp3, mp4, m4a, wav, and webm, up to 25 MB. Transcription runs on Gumloop’s servers using OpenAI transcription models (Whisper and GPT-4o Transcribe), so your agent only ever receives the text transcript, never the raw audio.

​
Message queue and steering

You do not have to wait for the agent to finish before sending your next message. The message queue lets you line up follow-ups while the agent works, and it picks them up between steps.
Message queue showing multiple queued messages in different states
Because queued messages are injected at natural breakpoints, they act as steering: redirect the agent (“focus on the Q3 numbers instead”), add missing context (“the deadline is Friday”), or stack a sequence of tasks. You can edit, reorder, or remove a queued message before it is delivered.

​
Context usage meter

The circular meter in the bottom-right of the chat input shows how much of the model’s context window is in use. Hover it for a breakdown across System, AI Instructions, Abilities, Tools, Skills, Subagents, and Conversation.
Context Usage Meter showing a token breakdown by category
If context fills up, reduce the tools or skills attached, switch to a model with a larger context window, or rely on auto summarization to compress older messages.

​
Incognito mode

Incognito conversations are not saved to the database. They are held in temporary memory and auto-deleted after 24 hours. Toggle Incognito from the chat input menu before sending.
BehaviorStandard chatIncognito chat
Message storageSaved permanentlyNot saved
Visible in history and searchYesNo
Included in data exportsYesNo
Files and artifactsStored permanentlyAuto-deleted after 24 hours
Used for reflectionsYesNo, excluded
Incognito applies to the whole conversation, including any subagents it spawns. Once it expires, messages and files are permanently gone.

​
Understanding credit costs

Agents consume credits based on AI model usage, tool calls, and any workflows they run. Cost depends on the model, message length, conversation history, and the number of tools available. Open Chat Details on any conversation to see the credit breakdown, split into Chat & Reasoning and Tool Calls, along with the model and source of the chat.
Chat Details panel showing source, model, participants, and a credit breakdown split into Chat and Reasoning and Tool Calls

Credits

Full model pricing, workflow and integration costs, and how to track usage.

​
Settings

The Settings tab covers how the agent presents and behaves outside its core logic.
Settings tab showing Personalization, Agent Details, Chat Preferences with Smart Suggestions and File Sharing Behavior, Slack Preferences, Secrets, and Danger Zone
SectionWhat it controls
PersonalizationThe agent’s icon, name, and description.
Agent DetailsMetadata about the agent, plus Make a Copy to duplicate it.
Chat PreferencesSmart Suggestions (suggested next actions in chat) and File Sharing Behavior (Default, Organization, or Anyone with the link) for files the agent generates.
Slack PreferencesHow the agent behaves in Slack, including thread responses and attribution.
SecretsEnvironment variables and secrets the agent can use. These are injected into the code sandbox at runtime and never logged.
Danger ZoneDelete the agent.

​
Deploying agents

An agent is not limited to the Gumloop chat. Deploy it where your team already works.

Slack

Deploy agents to Slack channels for team-wide access.

Microsoft Teams

Deploy agents to Microsoft Teams channels.

Email

Give your agent an inbox it can read and reply from.

Hosted pages

Share a public or private hosted chat page for your agent.

​
Finding agents

The Agents page lists every agent you can access. Use the tabs to switch views:
TabWhat it shows
MineAgents you created.
Shared with meAgents others shared with you directly or through your organization.
OrganizationAll agents visible to your organization.
Agents page showing the Shared with me tab with agent cards
Each card shows the agent name, connected apps, creator, and last activity. You can search by name and switch between grid and list views.

​
Managing chats

Every conversation appears in the sidebar. Right-click a chat or open its three-dot menu to manage it.
Chat context menu showing Share, Rename, and Delete options
ActionWhat it does
ShareShare the conversation as a read-only link.
RenameGive the chat a custom name so you can find it later.
DeletePermanently remove the conversation.
Rename chats to keep your sidebar organized. “Q2 Marketing Plan” is easier to find than an auto-generated title.

​
Guardrails

For agents that take real actions, layer on guardrails so they stay within bounds.

Human in the Loop

Make the agent pause for approval before write, delete, and other sensitive tool calls.

App Rules

Set conditions that block or flag specific tool calls at the agent or organization level.

​
Self-improving instructions

Your agent can update its own system prompt during a conversation. Correct it once (“always check Salesforce first,” “keep emails under 100 words”) and it edits its instructions so the same mistake does not happen again. Changes take effect on the next step and persist across future conversations. The toggle sits below the system prompt editor and is on by default; there is no version history, so revert by editing the prompt manually.

​
AI advanced settings

Click Advanced in Agent Preferences to fine-tune how the model behaves. This is an advanced area: the defaults are optimized for a good balance of performance, cost, and reliability, so most people never need to touch it.
AI Advanced Settings panel
Settings are organized into three tabs.
Per-provider model parameters. Settings are stored per provider, so switching models preserves your preferences for each.
The Model tab of AI Advanced Settings showing Max Steps, extended thinking, temperature, and max tokens
ParameterNotes
Max StepsHow many tool calls the agent can make before it has to respond. Default 100, range 1 to 200. Increase for tasks that chain many tools.
Reasoning / thinkingHow much the model deliberates before answering. OpenAI uses Reasoning Effort, Claude uses Extended Thinking with a token budget, Gemini uses Thinking Level.
TemperatureOutput randomness. Lower is more focused, higher is more creative. Claude forces this to 1.0 when extended thinking is on.
Max TokensUpper bound on generated tokens. Defaults to Auto.
Top P / Top KSampling controls. Adjust Temperature or Top P, not both.
Parallel tool callsWhether the model can call multiple tools at once. Disable for strict sequential execution.

AI Models

Browse the full model catalog by tier, vision models, and bring-your-own-key (BYOK).

​
Best practices and troubleshooting

Begin with two or three apps and short instructions. Test, watch how the agent behaves, then add tools and rules based on real usage. Avoid launching with 15 tools and a 2,000-word prompt.
When the agent makes a mistake, ask it: “What could I add to your instructions to prevent this?” Then have it update its own prompt or a skill. Review conversation history for patterns.
Be explicit about what the agent should never do without approval: delete records, send emails, make purchases, or modify production data. Back this up with Tool Management approvals.

​
Next steps

Agent Triggers

Run agents automatically on a schedule or in response to events.

Agent Skills

Build reusable knowledge packs that teach agents how to do specific work.

Evaluations

Grade agent responses and catch regressions before they ship.

Reflections

Let agents review their own work and propose improvements.

Code Sandbox & Secrets

Run code securely and manage agent secrets.

Agent Node

Embed agents in workflows for chaining and batch processing.