The chat
Conversations are multi-turn and saved. You can rename a chat, archive it, restore it later, and stop a turn while it’s still running. The assistant works from the same read-only data tools as the MCP server, but inside FixAEO. It can read the actual answer text an engine returned, and it can scrape a cited page to check what’s really on it. When it makes a claim, it can show you where the claim came from. A few controls sit around each turn:- Attach context. Paste or add material you want this turn to work from.
- Pick its tools. Choose which of its data tools it’s allowed to use for a turn, so you can keep an answer narrow.
- Get a deliverable. Some turns produce a downloadable file, such as a brief or a draft. You can preview it before you download it.
Agents
The second sidebar entry, Agents, opens the agent builder at/app/workflows/. Where the chat answers one question at a time, an agent runs a fixed set of steps end to end. You can build one by describing it in plain language or on a visual canvas, start from a ready-made template, save reusable skills and resources, and schedule or automate runs.
Agents have their own section in these docs:
Agents
Build, test, publish, and run agents over your own data.
Templates
Ready-made recipes and prebuilt agents to start from.
Skills
Reusable instructions you attach to any agent.
Scheduling and automation
Run agents on a schedule, by webhook, or in bulk.
The chat and its drafts are grounded in numbers you can verify elsewhere in FixAEO. Letting an agent run steps on its own is newer, so treat what an agent produces as a draft to review, not an answer to trust unattended. Read the evidence it cites before you act on it.
Nothing goes live on its own
Every result is a draft. It waits at a review step for a person to approve or reject it. FixAEO does not publish to your site, send email, or write to your CMS on its own. When an agent finishes, you get the output and the evidence behind it, and the next move is yours. This is a deliberate boundary, not a missing feature. The work is done for you; the decision stays with you.Who can use it
The AI Marketer and agents are on paid plans. The Free plan doesn’t include them. If you try to use them there, FixAEO points you to upgrade, and access is enforced on the server, not just hidden in the interface. The work runs on credits. Each plan includes a monthly allowance: Lite includes 100, Growth includes 400, and Enterprise is custom. Credits meter the measured cost of the AI work, so a heavier task costs more than a light one. Your remaining balance for the cycle shows in Settings → Usage. On a team, AI Marketer credits are a single shared pool for the whole workspace. The Usage page attributes spend back to each member so you can see where it’s going.Don’t confuse the AI Marketer and Agents with the Agent Analytics area (Crawl insights and Crawlability). Those read your server logs to show which AI crawlers fetched which pages. They’re about crawler traffic, not about generating or reviewing marketing work. See Agent analytics.
Related
MCP server
The same read-only data tools, pointed at your own ChatGPT, Claude, or Cursor.
Plans and limits
Which plans include the AI Marketer, and the credit allowance for each.
Improve
The checklist of recommendations you work through and mark off.
Agent analytics
A different feature: what AI crawlers fetched, read from your server logs.