Research where you already work
Keywords, live Google results, and visibility stay in the same chat that ships the site. You do not export CSVs into another tab.
Fig gives your agent live keywords, Google results, and visibility research. You approve sign-in and Checkout. Prepaid from $10.
Give this to your agent
Set up Fig - https://figaeo.com/llms.txt
Live Google results, keywords, visibility, backlinks, and local lookups — asked in chat, not in a suite. Copy the setup prompt, connect, then research.
Keywords, live Google results, and visibility stay in the same chat that ships the site. You do not export CSVs into another tab.
Your agent calls hosted lookups. You are not maintaining keys, proxies, and parsers just to fetch search data.
About 4,000 live SERPs (top 10), or about 500 keyword idea lookups at the default of 20 ideas, per $20 when the pack is spent entirely on either operation. A $20 pack is enough headroom for a focused pass on one site.
Fig is research your agent can call. Connect the agent you already use, approve sign-in once, then ask in chat.
Paste the setup prompt into chat. You approve GitHub or Google sign-in. Then ask for research.
Give this to your agent
Set up Fig - https://figaeo.com/llms.txt
{
"mcpServers": {
"figaeo": {
"url": "https://figaeo.com/mcp"
}
}
}
{
"name": "Fig",
"url": "https://figaeo.com/mcp",
"transport": "Streamable HTTP",
"auth": "MCP OAuth 2.1"
}
claude mcp add --transport http figaeo https://figaeo.com/mcp
codex mcp add figaeo --url https://figaeo.com/mcp
openclaw mcp add figaeo --url https://figaeo.com/mcp --transport streamable-http --auth oauth
mcp_servers:
figaeo:
url: "https://figaeo.com/mcp"
auth: oauthgemini mcp add --transport http figaeo https://figaeo.com/mcp
Install Fig as a hosted Streamable HTTP MCP server at https://figaeo.com/mcp. Also install the Fig SEO/AEO router and specialist skills from https://figaeo.com/skills/fig-seo-aeo/SKILL.md, https://figaeo.com/skills/fig-site-audit/SKILL.md, https://figaeo.com/skills/fig-search-research/SKILL.md, https://figaeo.com/skills/fig-answer-research/SKILL.md, https://figaeo.com/skills/fig-content-strategy/SKILL.md, https://figaeo.com/skills/fig-content-validation/SKILL.md into your native skill folders. A tools-only connection is incomplete. Open the connect card or browser and ask the User to complete MCP OAuth 2.1 with GitHub or Google. At $0 remaining, call create_checkout (default $20, minimum $10), give the returned short figaeo.com/pay link to the User to pay on their own computer, and wait for Checkout to finish. Then retry the same tool. Then research a product site: call research_domain for the site, suggest_keywords from the strongest terms, and get_keyword_metrics on a short list such as hosted MCP server. Prefer compact format. United States and English are the defaults unless the User names another market. After setup or skill migration, do not stop at installation. Inspect the current project and identify its target site. Run a small baseline with fig-site-audit and the relevant read-only Fig research tools before recommending next steps. Show the User a concise bullet list of findings, then offer these paths: - Fix technical SEO, crawlability, or indexability. - Find search demand, intent, and SERP opportunities. - Improve AI-answer visibility, citability, or access. - Build a content strategy or content brief. - Implement and validate a specific content or site change. Ask what they would like to pursue, or invite another goal. Route the chosen work through fig-seo-aeo.
Prepaid credits, not a monthly plan. Your agent spends them on lookups. Top up when remaining credits cannot cover the next lookup.
Recommended pack
$20
prepaid credits
Enough to research, act on, and recheck one website for about four weeks.
Minimum $10. The $20 pack is the recommended default.
The $10 minimum and $20 recommendation are credit subtotals. Applicable tax is added at Checkout.
Deeper result pages and site: queries cost more.
Give this to your agent
Set up Fig - https://figaeo.com/llms.txt
Counts assume the whole pack went to that one lookup at its default. Mixing lookups spends the same prepaid balance.
| Operation | Calls |
|---|---|
| Live SERPsGoogle organic, top 10 | 4,000 |
| Keyword idea lookups20 ideas each | 500 |
| Keyword metrics taskslive Google Ads metrics, not idea lookups | 110 |
| Domain keyword lookups20 keywords each | 500 |
| SERP competitor lookups20 competitors each | 500 |
| Backlink overviewsreferring domain summary | 400 |
| Backlink lists20 rows each | 400 |
| Maps lookupslive Google Maps, 100 results | 4,000 |
| Business listing lookups20 listings each | 500 |
| AI mention lookups20 mentions each | 80 |
Fig returns live SEO and AEO research: keywords, search results, competitors, backlinks, local listings, and AI visibility. You and your agent decide what to change. Fig does not edit your site. There are no ranking guarantees.
$20 is the recommended pack: about 4,000 live SERPs (top 10) or about 500 keyword idea lookups (20 ideas each) when spent entirely on either operation. That is enough to research, act on, and recheck one website for about four weeks. Deeper SERPs and site: queries cost more. A typical focused pass is still about 20 idea lookups and about 120 SERPs per two weeks (about 30 keywords, 3 snapshots). A $20 pack is about four weeks of that headroom, not a spend target.
Minimum $10. The recommended pack is $20. The $10 minimum and $20 recommendation are credit subtotals. Applicable tax is added at Checkout.
Give your agent the setup prompt, then sign in with GitHub or Google. That email is your Fig account. Your agent starts Checkout and shares a short figaeo.com/pay link. Pay on your own computer. On Grok, you can finish with Stripe Link after you connect Link. After you pay, ask the agent to retry the same lookup.
Yes. One Fig account follows the verified email from sign-in, so remaining credits are shared across every agent that uses that email.
When remaining credits cannot cover the next lookup, your agent starts Checkout and shares a short figaeo.com/pay link. Pay on your computer, then ask it to retry the same research.