Fimo
Fimo is an autonomous website platform that takes over once a coding agent has built your site: a visual CMS on your real code, scheduled AI agents and managed hosting, for developers, marketers and founders.
What is Fimo?
Fimo describes itself as an autonomous website platform, or an AI CMS: the layer that takes over once a coding agent has built your website. It does not generate the site, it operates it. The product is organized in three stages, Build to compose the site, Manage to orchestrate agents and content, and Serve to deliver it to users.
The entry point is your own repository. Fimo works alongside Claude Code, Codex, Cursor, Devin Desktop, GitHub Copilot, Warp and Antigravity, and supports React Router, Vite with React, Next.js, Astro and SvelteKit. Nothing moves into a proprietary builder: the site stays where your developers left it, and every edit lands as a commit.
Three surfaces sit on top of that code. A real-time visual editor lets anyone click an element on the page and change it, with the underlying code updated automatically. Fimo Studio is the dashboard non-developers use for content, assets, agents and publishing, with an AI chat that produces pages, content, images and fixes from a plain-language brief. The fimo CLI covers content, media, forms, deployments, locales and webhooks.
The other half of the product is agents. An agent is a folder of plain files in the repository: GOAL.md is required, alongside optional SOUL.md, capabilities.yaml, config.yaml and folders for skills, scripts, references and assets. Built-in skills cover deployment, content and schema editing, SEO, translation and media generation. Each agent picks its own model among Claude, GPT, Gemini, Grok and Llama, writes files, and keeps its own memory between runs. Each works in an isolated environment with a dedicated server, database and asset bucket, and configurable merge rules decide what auto-merges, such as content and translations, and what waits for human approval, such as schema changes and deletions. Every page is also rendered in Markdown for AI crawlers and LLMs.
Managed hosting is included on every plan, with custom domains on the paid tier, and migration tooling covers Webflow, Framer, Squarespace, Wix, Carrd, static exports and code repositories. Fimo reports 4,800+ sites launched, 15,000+ agents created, 171,000+ runs and 2,000+ teams. It comes from the founders of Strapi, Pierre Burgy, Aurélien Georget and Alexandre Bodin, but unlike Strapi it is not open source.
What it does
- Makes a website built by a coding agent editable by the whole team, directly on the real code
- Runs scheduled agents for SEO, translation, content, Lighthouse and accessibility on the live site
- Deploys the site and its agents together, with managed hosting, SSL, CDN and custom domains
- Isolates every change in a dedicated environment with its own database and asset bucket
- Returns each modification as a reviewable, revertible commit and pull request
When to use Fimo / When not to
A quick filter to help you decide if Fimo is the right fit.
When to use Fimo
- Teams whose website was built by a coding agent such as Claude Code, Codex or Cursor, and now has to be kept alive after launch
- Freelance developers and agencies that deliver a site and want to stop receiving copy-change requests forever after
- Marketing teams that need to edit live pages and publish content without going through a developer
- Owners of multilingual sites, who can hand translation and locale management to scheduled agents
- Solo founders and growth teams with heavy SEO and GEO needs, who want to keep owning the code while SEO agents work on it
When not to use Fimo
- Anyone looking for a tool to build a website from scratch: Fimo states it is not a page builder and only steps in once the site exists
- Teams with an open-source requirement, since Fimo itself is not open source, unlike Strapi
- Projects that cannot live in a Git repository, because everything goes through real versioned code, commits and pull requests
- Workloads carrying health data covered by HIPAA, financial account numbers, government identifiers or biometrics, all forbidden by the terms
- Users under 18 under the terms, people in territories under United States sanctions, and teams that need a contractual SLA or advanced security without an Enterprise budget
How to use Fimo
A typical end-to-end flow, from setup to results.
- Start from a website that already exists as code in a Git repository
- Install Fimo with fimo init: an SDK wires up the existing content to make it editable, with no rebuild and no migration
- Connect your coding agent through the remote MCP server at https://api.fimo.ai/api/mcp, authenticated with OAuth 2.1 and PKCE, with no token to paste
- Set up your organization: every project belongs to one, with roles and permissions assigned per person, agents included
- Hand editing to the team, so marketing changes content on its own in Studio without a developer
- Create agents by describing what they should do in plain language and choosing a cadence, daily, weekly, on every deployment, on content publication or on demand; Fimo then writes the agent as code in the repository
- Let agents work in isolated environments and open pull requests, with merge rules deciding what needs human approval
- Manage languages with fimo locales list, fimo locales add and fimo locales set-default, or through Settings then Languages in Studio
- Review each change on its own preview URL before it goes live
- Ship the site and its agents together with a single fimo deploy, and roll back in one click if needed
Pros & Cons
Pros
- The code remains yours, in your own GitHub repository, with no proprietary builder and no lock-in
- Every change is a versioned commit that can be reverted in one click
- Agents work in isolated environments and never touch production without a merge
- The terms categorically exclude training models on customer data and outputs
- Pricing is charged to the organization rather than per seat, with unlimited users on the paid plan
- Managed hosting, SSL and CDN are included from the free plan onwards
- Substantial public documentation, a full CLI, an MCP server, a published subprocessor list and a public DPA with standard contractual clauses, from a team that came out of Strapi and a decade of content infrastructure
Cons
- Young product: the About page dates Fimo's launch to 2026 and the domain was only registered in 2024, so there is little track record to lean on
- Not open source, unlike Strapi, whose founders built it
- It assumes a Git repository and a code toolchain; without a developer at the start there is no entry point
- The terms grant the publisher a worldwide, perpetual license over Customer Data other than personal information, and the publisher owns Usage Data outright and may exploit it for any business purpose
- Credits do not roll over, daily credits expire each day, nothing is refundable, and all credits are forfeited if the account is terminated for breach
- No security certification is published: neither SOC 2 nor ISO 27001 appears anywhere on the site, and no list of interface languages is published either
- The published documents contradict one another in places: the minimum age is 18 in the terms and 16 in the privacy policy, the Article 27 EU representative carries a San Francisco address, the agents page names Pro and Business plans that the pricing page does not list, and there is no contact page, the Book a demo button exposing no destination and support running through email and Discord
Pricing & Plans
Fimo operates on a freemium model. A permanent free plan is available at no cost, offering 2,500 AI credits per month, 1,000 free daily credits capped at 5,000 per month, 5 GB of bandwidth, up to 3 projects and one active user per project. The lowest paid entry point is the Premium plan at USD 25.00 per month, billed to the organization rather than per seat or per project, and raising the allowance to 20,000 AI credits per month, 1,000 daily credits capped at 30,000 per month, 100 GB of bandwidth and unlimited projects and users. An Enterprise plan is priced on quotation. Under section 4 of the terms, paid plans are billed monthly in advance, renew automatically, and fees are non-refundable except where the law provides otherwise; credits do not carry over from one month to the next, are non-refundable and non-transferable, and hold no monetary value. Payment is taken through Stripe or another designated provider, with Hyperline cited for metering. The agents page additionally mentions billing at USD 0.10 per environment-hour beyond the included allowance.
- 2
- 500 AI credits per month
- 5 GB of bandwidth per month
- up to 3 projects
- 1 active user per project
- 14-day version history
- 7-day analytics retention
- managed hosting with SSL and CDN
- 20
- 000 AI credits per month
- 100 GB of bandwidth per month
- unlimited public and private projects
- unlimited active users
- 90-day version history
- 5-year analytics retention
- custom domains
- custom limits
- advanced security and compliance
- dedicated support
- custom SLA
Data, GDPR & hosting
A consolidated view of how Fimo handles your data.
GDPR overview
GDPR implementation is concrete and documented. A data processing addendum effective 21 January 2026 is published at /data-processing-addendum, drafted on Regulation 2016/679 and annexed with the EU standard contractual clauses covering transfers outside the EEA. The privacy policy, effective 4 August 2026, lists rights of access, rectification, objection to profiling, portability in CSV or JSON, erasure, and complaint to a supervisory authority; requests go to support@fimo.ai, with the subject line DSR for the EEA and the United Kingdom, and are answered within one month, extendable for complex cases. A subprocessor list is published and refreshed at least yearly, with a 20-business-day objection window; breaches are notified without undue delay; data is deleted within 30 business days of termination on written request. One reservation: the Article 27 representative is a Data Protection Officer carrying a San Francisco address, not an establishment in the Union.
Who owns the data?
Under the terms effective 21 January 2026, you own your Customer Data, meaning the applications, websites and projects you build, along with the Output the service generates; the site lives in your own GitHub repository, and Fimo advertises no lock-in and no export button to wait for. Strapi, Inc. retains ownership of the platform, software, models and algorithms (section 11.A) and owns Usage Data such as telemetry, logs and metrics, which it may exploit for any business purpose (section 11.C). Section 7.C grants the publisher a worldwide, perpetual, royalty-free license over Customer Data other than personal information, for operation, improvement, benchmarking and analysis. Section 7.D forbids selling personal information or sharing it for advertising.
Reuse rights
You may reuse your own content and the generated Output freely and without asking permission: you own both, the code sits in your repository, and no proprietary format stands in the way. On the platform side, the privacy policy effective 4 August 2026 records the collection of contact details, payment information, IP address, login, browser, time zone and geolocation, product usage journeys, and visitor data from your published site. These serve four purposes: running the service (contract and legitimate interest), improving it through testing, heatmaps, analysis, profiling and machine learning, providing customer support (contract), and sending newsletters and invitations (consent only). Sensitive data is not collected unless you consent specifically or the law requires it. Section 7.B of the terms confirms that inputs and prompts are passed to AI systems supplied by Fimo, by third parties such as OpenAI, Google and OpenRouter, or by models you connect yourself. Section 7.C is explicit: Fimo does not use Customer Data or Output to train, fine-tune or otherwise improve the machine learning models or AI models used by the service. Aggregated, anonymized and de-identified data may still feed product improvement, benchmarking and new product development. Lemlist is used to identify visitors to the public website and capture intent signals.
Data retention & training
Hosting summary
The privacy policy states that Fimo is a product of Strapi, a United States Delaware corporation, with primary storage of your information in the United States and the EEA. Processing may also take place in other countries where Strapi has operations, employees or subprocessors. Section 9 of the data processing addendum, effective 21 January 2026, confirms that processing can occur outside the protected area, including in the United States, and relies on the EU standard contractual clauses to cover those transfers. The infrastructure named by Fimo includes Vercel, Cloudflare, Databricks, Blaxel and Relace, while the terms name OpenAI, Google and OpenRouter as third-party AI providers. A subprocessor list is published on a Notion page and updated at least once a year, and the addendum points to a security page at trust.strapi.io. At product level, each agent environment gets its own server, database and asset bucket, so a running agent stays isolated from production. No SOC 2 or ISO 27001 certification is published anywhere on the site.
Where Fimo works
Country-level availability.
Not available in
Things to keep in mind
Risks and trade-offs to weigh before adopting Fimo.
- Data ownership: the terms grant Strapi, Inc. a worldwide, perpetual, royalty-free license over Customer Data other than personal information, including for benchmarking and analysis, while Usage Data belongs outright to the publisher, which may exploit it for any business purpose
- Credit economics: monthly credits do not roll over, daily credits expire each day, nothing is refunded, and every credit is forfeited if the account is terminated for breach of the terms
- A site served on a fimo.com subdomain can be reclaimed by the publisher with only seven days' notice and no compensation, so a custom domain is the only durable address
- Forbidden data: health information covered by HIPAA, financial account numbers, government identifiers and biometrics may not be placed in the service at all, and the burden of keeping them out falls on the customer
- Autonomy invites complacency: the terms disclaim any warranty of accuracy for AI output and forbid unreviewed use in critical contexts, yet agents that publish on their own schedule make it easy to stop reading what goes live, and an unsupervised SEO or translation agent ships its mistakes at the same cadence as its successes
- Legal recourse is narrow: mandatory JAMS arbitration, waiver of class action and jury trial, and California law; infrastructure providers may also exploit aggregated usage data on their own account
- The published documents contradict one another on points worth checking before signing: the minimum age is 18 in the terms and 16 in the privacy policy, and the representative appointed under Article 27 of the GDPR carries a United States address rather than an establishment in the Union
Setup & Integrations
Technical difficulty
Moderate, and front-loaded on the developer side. The prerequisite is a site that already exists as code in a repository; installation runs through an SDK that wires up existing content, with no rebuild and no migration, starting with fimo init in the CLI. Migration tooling covers Webflow, Framer, Squarespace, Wix, Carrd and static exports. After that, Studio lets non-developers edit content and steer agents without touching code, and an agent can be created in plain language. Advanced customization is another matter: YAML files, scripts, and non-interactive operations that must pass the --yes and --json flags.
Deployment
Integrations
Behind Fimo
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Fimo.
Frequently asked questions
What is an autonomous website?
How is Fimo different from an AI website builder?
Who owns the code produced with Fimo?
Do I need to be a developer to use Fimo?
Can I keep using my AI coding tools?
What do the agents actually do?
Is there a free plan, and what does the first paid tier cost?
Is my data used to train AI models?
Where is my data hosted, and is a DPA available?
Is there a minimum age?
Should you pick Fimo?
Fimo is not a website builder, and it is careful to say so. It is the operating layer that comes after one: a visual CMS mounted on your real code, a fleet of scheduled agents, and managed hosting, all wired to the repository your coding agent already filled. That positioning is the single best filter for the decision. If your site already lives as code in a Git repository, Fimo is a coherent answer to the question of what happens after launch. Marketing edits without a developer, SEO and translation agents run on their own schedule, and every change arrives as a reviewable commit that reverts in one click.
The strong points are unusually concrete. You keep the code, in your own repository, with no proprietary format to escape from. The terms categorically exclude training models on customer data and outputs. Pricing is charged to the organization rather than per seat, and hosting, SSL and CDN are included from the free plan.
The reservations deserve equal weight. The product is young. The terms grant a worldwide, perpetual license over customer data other than personal information, and hand usage data outright to the publisher. Credits expire and are never refunded. No SOC 2 or ISO 27001 certification is published. And the legal pages contradict one another in places: a minimum age of 18 in the terms against 16 in the privacy policy, and a representative appointed under Article 27 of the GDPR given a San Francisco address rather than an establishment in the Union.
Worth a serious look for a code-first team that has shipped a site and now has to run it. Irrelevant to anyone still looking for something to build that site with.
- Choosing a selection results in a full page refresh.
- Opens in a new window.