Custom agents

Build agents for anything
your business needs.

Dozens of pre-built agent templates for common tasks. Custom prompts, tools, and behaviors. Define exactly how agents work, what they can access, and how they collaborate. Your business, your agents.

In Action

Create a custom agent in seconds

Define what it does, how it behaves, and which tools it can use.

mode | custom agent builder
$ mode agent create brand-voice
[builder] Setting agent identity ···
[tools] Attaching web_search, file_read, style_check ···
[prompts] Loading brand guidelines as context ···
[validator] Testing agent behavior ···
[builder] Agent "brand-voice" registered
[tools] 3 tools attached
[prompts] System prompt loaded (2.4k tokens)
[validator] 8/8 behavior checks passed
✓ Ready to launch.
Run with: mode --agent brand-voice
$
Catalog

Pre-built agents, ready to deploy

Start with a proven agent template. Customize it to fit your exact workflow.

Content

Brand Voice Agent

Enforces your tone, terminology, and style guidelines across every piece of content. Reviews drafts and rewrites off-brand copy automatically.

Legal

Compliance Checker

Scans documents, code, and processes against regulatory frameworks. Flags violations and suggests fixes for GDPR, SOC 2, HIPAA, and more.

Research

Domain Researcher

Deep-dives into any industry vertical. Gathers market data, competitor intel, and emerging trends, then delivers structured reports with cited sources.

Support

Customer Support Agent

Handles tickets, resolves common issues, escalates edge cases. Learns your product inside and out from your docs and past conversations.

Analytics

Data Analyst Agent

Queries databases, builds visualizations, and surfaces insights from your data. Answers business questions in plain language with the numbers to back it up.

Engineering

Code Reviewer

Reviews pull requests for correctness, security vulnerabilities, performance issues, and style violations. Catches what linters miss.

Builder

Build your own from scratch

Define custom prompts, attach tools, set behaviors, and deploy agents tailored to your exact needs.

1. Define the Agent

Give your agent a name, a system prompt, and a role. Tell it what it's responsible for, what it should prioritize, and how it should communicate. Every agent gets a custom identity that shapes its behavior.

2. Attach Tools

Choose from built-in tools like web search, file operations, API calls, database queries, and shell commands. Or write your own custom tools. Each agent only gets access to what it needs.

3. Set Behaviors

Configure guardrails, output formats, and escalation rules. Define when the agent should ask for approval, retry, or hand off to another agent. Full control over how autonomous it gets.

4. Deploy & Orchestrate

Run your custom agent solo or wire it into a multi-agent workflow. Agents share context, pass results to each other, and work in parallel. One command launches the whole team.

Categories

Agents for every department

Whatever your team does, there's an agent type for it.

Engineering

Engineering & DevOps

Code reviewers, QA testers, security auditors, CI/CD monitors, dependency updaters, migration assistants, and infrastructure agents.

Content

Content & Marketing

Brand voice enforcers, content moderators, SEO optimizers, social media schedulers, copywriters, and campaign analysts.

Operations

Operations & HR

Onboarding assistants, policy enforcers, scheduling agents, workflow automators, document processors, and knowledge base curators.

Security

Security & Compliance

Security auditors, compliance checkers, vulnerability scanners, access reviewers, incident responders, and audit trail generators.

Support

Success & Support

Customer support agents, lead qualifiers, proposal writers, CRM updaters, churn predictors, and account health monitors.

Analytics

Data & Research

Data analysts, domain researchers, trend spotters, report generators, survey processors, and competitive intelligence agents.

Orchestration

How agents work together

The real power is not one agent. It's dozens of specialized agents running in parallel, sharing context, and building on each other's work.

Parallel Execution

Independent agents run simultaneously, not sequentially. A compliance checker, a brand voice reviewer, and a data analyst can all work on the same project at the same time. Results in minutes instead of hours.

Shared Context

Agents share a common context layer so each one knows what the others have found. No duplicated work, no contradictory outputs. The compliance agent sees what the researcher discovered and builds on it.

Dependency Graphs

Mode Agent understands which tasks depend on others. It runs everything possible in parallel and queues dependent work automatically. You describe the goal and the Agent figures out the execution order.

Mix and Match

Combine pre-built agents with your custom agents in the same workflow. A research agent feeds a writer agent which feeds a compliance agent. Build pipelines that match how your business actually operates.

Ready to build your own agents?

Get up and running in minutes.