Every part of organic growth, run by an agent
Launch a prebuilt Slate agent that researches, writes, optimizes, and publishes across your stack — or build your own.
Most popular agents
The six teams turn on first.
SEO & AEO
Agents that audit, research, brief, write, and optimize for Google and for LLMs.
Content production
Agents that draft, illustrate, repurpose, and ship every asset to your CMS.
Nothing here fits? Build the agent.
Every agent above is open. Fork one, rewire it, or start from a blank canvas.
- Visual builder — loops, branches, code, and web search in any step
- 24+ models, picked per step
- Brand Kit, Skills, and Memory enforced on every run
- Publish to Webflow and WordPress, pull from GSC, GA4, Semrush, Ahrefs
Frequently Asked Questions
What are Slate agents?
Slate agents are autonomous AI workers that execute repeatable SEO, content, and growth tasks end-to-end. Give one a goal in plain language "refresh my 20 lowest-performing blog posts" or "find every Reddit thread where our product is mentioned" and the agent figures out the tool calls, the sequence, and the output. Unlike chatbots that only generate text, Slate agents can research, analyze, write, publish, and report across the full organic growth loop.
How are agents different from workflows?
Workflows follow a fixed path you define step by step. Agents decide their own path based on the goal you give them. A workflow is deterministic it runs the same way every time. An agent adapts: if a data source is unavailable, it tries another route. If the first approach doesn't produce useful results, it adjusts. Most teams start with pre-built agents and graduate to custom workflows when they need a repeatable, auditable process for a specific operation.
Do I need to know how to code to build or use an agent?
No. Every agent on this page is ready to run as-is. Fork one, adjust its instructions in natural language, and it adapts. If you want to build something from scratch, you describe the outcome in plain language no Python, no API keys to manage, no infrastructure to maintain. Slate handles the execution layer so your team focuses on strategy and review.
What can Slate agents actually do?
SEO & AEO keyword gap research, competitor page audits, domain-wide SEO audits, backlink analysis, schema generation, internal linking optimization, and content refresh scheduling.
Content production brief-to-article writing, AI content humanization, blog-to-CMS publishing, multi-channel content repurposing, and Reddit-to-blog adaptation.
Market intelligence competitor intelligence gathering, G2 review extraction and analysis, Reddit mention tracking, and content performance monitoring.
Sales & GTM prospect research and outreach planning, ICP and sales strategy builder.
Every agent is open. You can fork any of them, rewire the instructions, or start from a blank canvas.
How does quality control work when an agent produces content?
Agents draft, research, and assemble they don't publish without review unless you explicitly configure auto-publish. Every agent output passes through Slate's editor before going live. Track changes are visible. Tone and brand alignment are checked against your BrandKit. Teams set their own review gates: some approve every piece, others batch-review weekly. The agent handles the production volume; humans keep the quality bar.
Can I customize the pre-built agents?
Yes. Every agent on this page is a starting point, not a locked template. Fork any agent and edit its instructions, connected tools, output format, or approval rules. The agent builder lets you add or remove capabilities, connect different data sources, and set custom triggers. Most teams start with a pre-built agent, run it on a real task, then refine the instructions based on what they see the agent even writes its own improved prompt after a successful run.
How do Slate agents connect to my existing tools?
Agents connect to your data sources and publishing destinations through Slate's integration layer. Native connectors include Ahrefs, Semrush, Google Search Console, GA4, Webflow, WordPress, and Slack. For anything beyond that, agents can use MCP servers or custom API calls. You configure access once per connector, and every agent that needs that tool uses it without separate setup.
What's the difference between Cowork and a custom agent?
Coworker is Slate's built-in AI assistant you describe a task in chat and it executes immediately, using the tools and data available in your workspace. A custom agent is a saved, reusable configuration you can name, share, trigger on a schedule, and refine over time. Think of Coworker as an on-demand executor and custom agents as packaged, repeatable processes your whole team can run.
How quickly can I start using these agents?
Immediately. Pick any agent from the hub, click to open it, and give it a task. No setup, no configuration, no onboarding flow. The agent uses the data sources already connected to your workspace. If you're new to Slate, the free trial includes full access to every agent you can run real tasks from day one.
How many agents can I run at once?
There's no hard limit on concurrent agents. Each agent runs independently, so you can have a content refresh agent working on blog updates, a competitor intelligence agent monitoring rival pages, and a Reddit tracker agent surfacing mentions all in parallel. Slate's credit system meters execution, not concurrency, so the question is how much work you want to automate, not how many agents you can have.