# Jam > Jam is the most proactive AEO tool. It tracks what AI says about your company, researches what to fix, and does the work, from new pages to pull requests. Jam is an answer engine optimization (AEO) tool, also called GEO. It checks the answers that ChatGPT, Gemini, Perplexity, Google AI Overviews, Google AI Mode and Copilot Search (Bing) give to the questions your buyers ask, with Claude available as an add-on, and measures how often each one recommends you against your competitors and which sources it cites. Where most AEO tools stop at tracking, Jam keeps going. Research agents read the answers, open the pages they cite and check your site, then rank the changes most likely to get you recommended: new pages, edits to existing pages, technical fixes, outreach and community work. Jam then does the work. It writes pages in your voice, makes the fixes, and opens pull requests you review and merge. Teams without a repo get Markdown or copy-ready text. Nothing goes live until you approve or merge it. Jam is built for software companies first, but works for any company whose buyers ask AI what to use. You can run it from your own agent through the Jam MCP server (Claude, ChatGPT, Cursor) or the API. ## Pricing Jam plans start at $140 a month for 50 prompts scanned weekly on one AI engine, or $144 a month on all 6 default engines, with 1,000 work credits. The price rises with the prompts you track (up to 1,000), how often Jam scans, the engines you add and your monthly work credits, and every plan starts with a 14-day free trial. Enterprise plans above 1,000 prompts are custom. Full calculator: https://www.spreadjam.com/pricing ## Product - [Jam](https://www.spreadjam.com/): What Jam does, from tracking AI answers to shipping the fixes - [Tracking](https://www.spreadjam.com/product/tracking): How often each AI engine recommends you against competitors, with the sources every answer cites - [Research](https://www.spreadjam.com/product/research): Research agents that read the answers and the cited pages, then rank what to fix - [Actions](https://www.spreadjam.com/product/actions): Ranked actions Jam carries out: new pages, page edits and pull requests, or hand-offs to your own agent - [Integrations](https://www.spreadjam.com/product/integrations): The tools and repositories Jam connects to - [Agent experience](https://www.spreadjam.com/agent-experience): Run Jam from Claude, ChatGPT or Cursor with the MCP server and API - [Pricing](https://www.spreadjam.com/pricing): Plans priced by prompts, engines, scan cadence and work credits, each with a free trial - [How credits work](https://www.spreadjam.com/credits): What a work credit buys and the full price list - [Security](https://www.spreadjam.com/security): How Jam handles your data and access ## Evidence - [Case studies](https://www.spreadjam.com/case-studies): Customer prompt sets, measured AI visibility, and what moved after the changes shipped - [Changelog](https://www.spreadjam.com/changelog): Product updates, newest first ## Developers - [API reference](https://docs.spreadjam.com/api-reference/overview): Every endpoint of the Jam API - [Documentation](https://docs.spreadjam.com/): API, MCP server and product docs ## Company - [About Jam](https://www.spreadjam.com/about): What Jam is (AEO, GEO, AI visibility), how it works, pricing, the team, and how it differs from Jam.dev, JamAI Base and Jam Content - [Manifesto](https://www.spreadjam.com/manifesto): Why getting customers is now an engineering problem - [Our team](https://www.spreadjam.com/our-team): The people building Jam - [Investors](https://www.spreadjam.com/investors): The funds and founders backing Jam - [Careers](https://www.spreadjam.com/careers): Open roles ## Use Jam from an agent Jam's AI search visibility product (GEO) has a REST API and a remote MCP server, so a coding agent (Claude Code, Cursor, any MCP client or script) can do what a customer does in the dashboard: read visibility results and the AI answers behind them, read the Plan of recommended actions, track prompts, and act. - What it scans: ChatGPT, Gemini, Perplexity, Google AI Mode, Google AI Overviews and Copilot Search (Bing); Claude as an opt-in add-on. - API base: https://api.spreadjam.com/v1 (auth: `Authorization: Bearer sk_...`) - OpenAPI spec: https://api.spreadjam.com/openapi.json - MCP server (Streamable HTTP, same key): https://api.spreadjam.com/mcp Claude Code: `claude mcp add --transport http jam https://api.spreadjam.com/mcp --header "Authorization: Bearer sk_..."` - Claude.ai, Claude Desktop, ChatGPT and other OAuth MCP clients: add https://api.spreadjam.com/mcp as a custom connector and sign in; no key to paste (OAuth 2.1 with PKCE, metadata at https://api.spreadjam.com/.well-known/oauth-protected-resource). - No key yet? `POST https://api.spreadjam.com/v1/agent-connect` with optional `{"domain", "email", "agentName"}` (no auth). Give the human the returned `connectUrl` and `userCode` (the approval page shows the same code, so they can check the request is yours): they sign up, set up their site, choose a plan and approve. Poll `GET https://api.spreadjam.com/v1/agent-connect/{pollToken}` every `intervalSeconds` until `status` is `approved`; that response carries the API key. Store it: if it is lost before you use it, poll again with the same token within 10 minutes for a replacement. - Set Jam up for them: `GET https://api.spreadjam.com/v1/pricing` (no auth) lists plans and monthly prices (`priceTable` has every work credit level and the opt-in Claude add-on; credits change the price most); add `onboarding` (`site`, `brandName`, `description`, `audience`, `competitors`, `prompts`, `notConfusedWith`, `locale`, `plan`) to the agent-connect request and the link opens one review page where the human approves and starts the trial with one button; pending polls carry `stage` (what is left); an `awaiting_payment` poll adds `reason`, `resumeUrl` and `humanMessage` to send the founder, and an approved setup waits 24 hours for payment. Jam tracks exactly the prompts you send (fill free slots later with `POST /v1/prompts/suggestions`); give competitors a `domain` and any other names as `aliases`. A prompt that names a tracked brand (yours, or a competitor as in "X vs Y" or "X alternatives") is tracked but not scored; the response `warnings` lists each one, and `GET /v1/prompts` shows `scored` and `unscoredReason` per prompt. If another product shares the brand's name, send it as `notConfusedWith` (`{"name": "Jam", "domain": "jam.dev"}`) so answers about it are not counted as the brand; change it later with `PATCH /v1/brand`. - Existing customers can also create a key in Settings, Developer. - First calls: `GET /v1/me`, `GET /v1/analytics/overview`, `GET /v1/prompts?limit=20` (send `limit` on lists; each answers `pageInfo` with the `nextCursor` to pass back as `cursor`), `GET /v1/actions`, `GET /v1/credits`. - Docs for agents: https://docs.spreadjam.com - Sign up in a browser: https://www.spreadjam.com/start - Support: founders@spreadjam.com ## Guides - [Best Generative Engine Optimization (GEO) Tools (2026)](https://www.spreadjam.com/blogs/best-geo-platforms-developer-tools): GEO tools for developer-tool companies compared: Gauge, Profound, Peec, Otterly, AirOps and Jam, with evidence tests, a trial plan and a scorecard. - [Best AI Visibility Tools (2026) and How to Choose One](https://www.spreadjam.com/blogs/best-ai-visibility-tools): Peec, Otterly, AirOps, Profound, Gauge and Jam compared on engines covered, citations, published pricing and who does the work after the report. - [How to Correct Inaccurate AI Brand Descriptions](https://www.spreadjam.com/blogs/correct-inaccurate-ai-brand-descriptions): A practical method for verifying wrong product claims, inspecting cited sources, preparing factual corrections and measuring whether answers change. - [API Documentation for AI-Assisted Discovery](https://www.spreadjam.com/blogs/api-documentation-ai-discovery): Connect use-case guidance, constraints, API reference and complete examples so developers can evaluate and implement your product with the right context. - [How to Trace Sources Behind AI Recommendations](https://www.spreadjam.com/blogs/trace-ai-recommendation-sources): Map AI answer claims to attached URLs, inspect what each source supports and preserve redirects, canonical information, dates and unresolved provenance. - [Preserve Context in AI-Readable Developer Docs](https://www.spreadjam.com/blogs/developer-documentation-ai-context): Keep product, version, permission and outcome context intact when a developer or AI assistant reads a documentation section on its own. - [API Examples That AI Agents Can Use](https://www.spreadjam.com/blogs/api-examples-ai-agents): Build complete API examples with explicit prerequisites, authorization, response states, retry boundaries and executable local checks. - [GEO Agency vs GEO Software vs In-House: How to Choose](https://www.spreadjam.com/blogs/geo-software-agency-in-house): Should you buy GEO software, hire a GEO agency or build in-house? A responsibility matrix, comparable deliverables and illustrative workload costs. - [AI Citation Tracking at the URL Level](https://www.spreadjam.com/blogs/url-citation-tracking): Track the exact pages AI answers cite, keep redirects and query parameters intact, and calculate page-level citation frequency with clear denominators. - [Evaluate Sentiment and Accuracy in AI Answers](https://www.spreadjam.com/blogs/ai-answer-sentiment-accuracy): Review AI answers with separate sentiment, recommendation and factual-accuracy labels, worked difficult cases and an explicit reviewer-disagreement process. - [Choose Buyer Questions for Your GEO Program](https://www.spreadjam.com/blogs/geo-buyer-question-selection): Design a GEO question set around buyer tasks, constraints and decision stages, with separate branded diagnostics, clear exclusions and a stable change policy. - [Prioritize GEO Work Across Pages and Sources](https://www.spreadjam.com/blogs/prioritize-geo-improvements): Turn answer and citation evidence into a feasible GEO backlog with explicit priorities, owners, dependencies and verification, without speculative scoring. - [Measure Whether a GEO Content Change Helped](https://www.spreadjam.com/blogs/measure-geo-content-changes): Compare GEO content changes with a frozen question set, repeated observations, explicit denominators and a change log that preserves uncertainty about cause. - [API Reference, Use-Case Guide or Integration Tutorial?](https://www.spreadjam.com/blogs/api-documentation-page-types): Choose the right documentation page for evaluation, exact API behavior and guided implementation, with three worked examples for one fictional event service. - [Write Integration Tutorials with Clear Prerequisites](https://www.spreadjam.com/blogs/integration-tutorial-prerequisites): Build tutorials that work from a clean starting state, with explicit runtime, permissions, configuration, expected results, failure handling and teardown. - [Document API Authentication for AI-Assisted Work](https://www.spreadjam.com/blogs/api-authentication-documentation): Explain credential origin, permissions, client boundaries and lifecycle so developers can use API authentication examples without guessing the missing context. - [Document API Errors for Reliable Recovery](https://www.spreadjam.com/blogs/api-error-documentation): Explain API errors with status, domain problem, known side effects and safe recovery actions, supported by a runnable local policy simulator. - [Keep Deprecated API Examples out of Current Guidance](https://www.spreadjam.com/blogs/deprecated-api-examples): Maintain versioned API examples with clear lifecycle notices, explicit migration steps and tested contract boundaries, while keeping useful older docs. - [Write Useful Technical Product Comparisons](https://www.spreadjam.com/blogs/technical-product-comparisons): Compare developer products against the same task, constraints and evidence, with explicit failure boundaries and conclusions that change with the buyer's needs. - [Explain Where Your Developer Tool Fits in a Stack](https://www.spreadjam.com/blogs/developer-tool-stack-positioning): Show what runs where, what data crosses each boundary and who owns recovery, credentials and state so buyers can place your developer tool correctly. - [AI Visibility Playbook for API and SDK Companies](https://www.spreadjam.com/blogs/api-sdk-ai-visibility-playbook): Build one API and SDK evidence surface across package pages, runtime support, references, quickstarts and migrations; measure discovery and execution apart. - [AI Visibility for Integration Infrastructure](https://www.spreadjam.com/blogs/integration-infrastructure-ai-visibility): Build evidence-led pages for integration buyers around connector scope, credential ownership, sync freshness, tenant boundaries and recovery behavior. - [AI Visibility for Agent Infrastructure](https://www.spreadjam.com/blogs/agent-infrastructure-ai-visibility): Show agent-infrastructure buyers the task fit, tool contracts, permission boundaries and evaluation evidence needed to choose and use your product. - [AI Visibility for Developer Tools: A Playbook](https://www.spreadjam.com/blogs/ai-answer-engine-optimization): Build a developer-tool GEO program from buyer questions and exact sources to reviewed documentation changes, clear owners and repeatable measurement. - [How to Measure AI Search Visibility](https://www.spreadjam.com/blogs/ai-search-visibility-monitoring): How to measure AI search visibility: pick the question set, the answer-level metrics, citation evidence and the sample size that makes a baseline hold. - [Investigate Competitors in AI Recommendations](https://www.spreadjam.com/blogs/ai-search-ranking-optimization): Investigate which competitors appear in AI answers, which pages support their descriptions and what the evidence justifies changing on your own site. ## Legal - [Terms of Service](https://www.spreadjam.com/terms) - [Privacy Policy](https://www.spreadjam.com/privacy)