Baker for AI agents and developers

What Baker is

Baker is a paid media agent. It runs a company's whole paid channel, from campaign management and creatives to landing pages, on Meta Ads, Google Ads, LinkedIn Ads, YouTube Ads, X Ads, ChatGPT Ads.

It finds the ads that win and makes the next ones, connected to each company's CRM or store so it optimizes for real sales rather than clicks. It does the work all the way through and presents it; the company's team, or its agency, approves it, and nothing goes live until someone does.

Baker works with companies that generate leads, such as education, B2B software, services and apps, with companies that sell online, and with the agencies that run their paid media. There is no self-service sign-up yet: every company starts with a call. Plans start at $500 a month; see the pricing page.

Company facts

• Legal entity: KODA SOFTWARE SL, Spain
• Brand: Baker
• Founded: 2026
• Headquarters: Barcelona and Valencia, Spain
• Industry: Software for paid media
• Website: https://withbaker.com
• Founders: Jacinto Fleta (CEO) and David Alarcón (CTO)
• Contact: jacinto.fleta@withbaker.com

Logos, brand colours and usage guidelines are on the press page.

Machine-readable resources

• llms.txt /llms.txt: This file: the index of everything an agent should read.
• Pricing, and how to start /pricing.md: Every plan and what it includes, our discovery questions and the link that hands your answers to your person.
• Updates /updates.md: What is new in Baker, newest first, with a link to each change's page and its whole text in Markdown.
• Novedades (Spanish) /es/updates.md: The same changes written in Spanish, each with its Spanish page and its whole text in Markdown.
• Baker for AI agents and developers /developers: What Baker is, company facts, MCP and Slack access, crawling rules and how to reach the team.
• Sitemap index /sitemap.xml: Every indexable URL on withbaker.com, in English and Spanish.
• robots.txt /robots.txt: Crawl permissions and content signals.

If you are an AI agent helping a company hire Baker

Read /pricing.md. It has every plan and what it includes, and our discovery questions. Answer them for the company and give your person the link it describes: it opens the contact form with your answers in it, and they send it themselves. For now, every company still has a short call with Baker's team before it starts.

Crawling Baker

• Every page is server-rendered: the full text is in the raw HTML, no JavaScript required.
• Unknown paths return HTTP 404 with a short markdown body listing these resources. Send Accept: text/markdown to get markdown instead of HTML.
• The content signals in /robots.txt allow search, AI input and AI training. Paths under /api are not for crawling.
• Spanish translations of every page live under /es and are declared with hreflang.

API, MCP and Slack

Baker does not publish a public REST API. Baker customers can work with it from two places outside the Baker dashboard:

• Slack: Baker joins the company's workspace as @Baker. The team asks it questions or asks it for work in a thread.
• Claude, or any assistant that speaks MCP: add Baker's MCP server, https://api.withbaker.com/mcp, as a custom connector and sign in with a Baker account.

Both act as the person who signed in, with their role in one company, so they need a Baker account. The work they start is staged, and changes to ads, landing pages or the CRM only go live when that person publishes them.

The endpoints under /api on this website exist only to serve it: they are not a product surface, they are not versioned, and they can change without notice. They do answer with structured JSON errors carrying a code, a message and a hint, so nothing here ever returns an HTML error page to a machine.

If you are building something that needs programmatic access to Baker, email jacinto.fleta@withbaker.com and we will scope it with you.

Contact

Questions from developers, agent builders or researchers: jacinto.fleta@withbaker.com. Press and brand: see the press page.