fungi.computer Privacy Policy
Draft correction to Version 0.4 (October 25, 2026). Not effective until separately authorized publication.
Summary
We collect what we need to run your account, your Team, your Agents and their Computers, and to bill you. When your Agents work, what they read and write is sent, through Cloudflare AI Gateway, to our AI model providers, OpenAI and Moonshot AI. Don’t send anything you wouldn’t want a model provider to see. We don’t sell your information and we don’t use it for advertising. The signed-in product uses only the cookies it needs to keep you logged in. There’s no delete or export button yet. Email us to request deletion or a copy of your data. We have not yet verified that our process can delete or export every copy.
You won’t find complicated legal terms or long passages of unreadable text here. We’ve no desire to trick you into agreeing to something you might later regret.
This policy covers fungi.computer, my.fungi.computer and the Fungi API, run by Jamstack Consulting LLC (“Fungi”, “we”). You can reach us at privacy@fungi.computer or at 13005 West Flower Street, Avondale, Arizona 85392, United States.
Our policy covers
- Why we value your privacy
- How we collect information
- What information we hold, and where we store it
- What happens when AI works on your content
- What we use your information for
- Who has access to information about you
- Cookies
- How long we keep your information
- Your rights, and how to use them
- The steps we take to keep your information safe
- Where your information goes
- The launch list
- Children
- Who’s responsible for your information
- How to complain
- Changes to the policy
Why we value your privacy
Fungi runs AI Agents on your files, your code and your conversations. That’s a lot of trust to place in us, so we try to earn it:
- We collect only what we need to run the Service and bill you.
- Your prompts, and the content your Agents read, go to our AI model providers, OpenAI and Moonshot AI, through Cloudflare AI Gateway, so they can generate responses. We explain this below.
- We don’t sell your personal information, and we don’t use it for advertising.
- The signed-in product uses only necessary login cookies. There are no analytics in it.
- There’s no self-serve delete or export button yet. Email us to ask what can be deleted or exported and how completion will be checked.
How we collect information
- You give it to us when you sign up, set up a Team, invite people, pay, or email us.
- Your Agents create it when they work: conversations, transcripts, files and generated images.
- Our systems record it when you sign in and use the Service, for example which browser you use and the IP address it connects from.
- When enabled, the chat launcher on fungi.computer loads from our Hub when you visit. Opening it loads the conversation view. Messages you send are handled as Agent conversations and may be sent to our model providers as described below.
- Services you connect send it to us, for example your GitHub account ID when you sign in with GitHub.
What information we hold, and where we store it
| What | For example | Where we store it |
|---|---|---|
| Your account | Name, email and whether you’ve verified your email | A Postgres database on operator-managed servers |
| Sign-ins | Session tokens, the IP address and browser details (user agent) of each session, your list of devices | The same Postgres database |
| Sign in with GitHub | Your GitHub account ID and an encrypted login token | The same Postgres database |
| Teams | Team names, who’s in each Team and their role, invitations (including the email addresses of people invited), a log of membership changes | Databases on operator-managed servers |
| Billing records | Subscription status, number of seats, credit balance and history, who switched on auto-refill and when | Databases on operator-managed servers |
| Payment details | Card and billing details | Stripe only. We never see or store your full card number. When we create your customer record at Stripe, the only thing we give them is an internal Team ID. |
| Agent sessions | Your messages, Agent transcripts, the tools Agents used and what came back | Operator-managed servers |
| Computer workspaces | Files in a Computer, cloned repositories, files you upload, notes Agents write, and anything you paste in (for example your own API keys in a .env file) |
Virtual machines and storage on operator-managed servers |
| Apps and images | App source code, bundles, generated images and videos, and their prompts | App state and separately configured source, artifact and media stores. Deletion from one is not all-copy deletion. |
| Connected accounts | Login credentials for outside accounts you connect and AI provider keys you add | Hosted credential services, including Nango. Credentials copied into prompts, files or provider tools can also remain in Computers or transcripts. |
| Emails we send you | Your address and the message (sign-in codes, verification, receipts, service notices) | Cloudflare Email Service |
| Support and requests | Whatever you send us by email | Cloudflare Email Routing forwards it to our Fastmail inbox |
| Operational logs | Details of requests, errors, and (if the network allowlist ships) which websites a Computer tried to reach | Operator-managed servers, and Cloudflare for requests that pass its network |
| Backups, when configured | Copies of databases and stored files | Operator-selected backup storage. The off-site copy job encrypts before upload, but some local backup formats are plaintext. Configuration alone does not establish that backups have run. |
We don’t ask for sensitive information. Please don’t put other people’s sensitive information into Fungi unless you have the right to.
What happens when AI works on your content?
- What goes to our AI providers. When you or an Agent run a task, your prompts, the files and code the Agent reads, and the model’s answers are sent through Cloudflare AI Gateway to whichever model is handling the request: OpenAI, L.L.C. (GPT-6 Luna, GPT-6 Sol) or Moonshot AI (Kimi K3).
- How long they keep it. OpenAI’s API data controls distinguish abuse-monitoring logs from feature-specific application state. The usual up-to-30-day abuse-monitoring period has exceptions and is not a universal deletion deadline. We haven’t verified a retention limit for Moonshot AI; do not assume it deletes this data within 30 days.
- What they may do with it. OpenAI’s terms say it doesn’t train on API data by default. Moonshot AI’s terms don’t make us the same promise, so assume it may use this data to improve its models. Don’t send anything you wouldn’t want a model provider to see. That includes passwords, API keys and other people’s personal information.
- Images and video. Prompts for generating images are processed either by Cloudflare Workers AI, using models from Black Forest Labs (Flux) and Leonardo that run on Cloudflare’s own network, or, for GPT Image 2, by OpenAI through Cloudflare AI Gateway. Prompts for generating video, and any images given to the model, go through Cloudflare AI Gateway to Google (Veo), ByteDance (Seedance) or MiniMax, depending on the model. We haven’t confirmed how long these video providers keep data or whether they use it to improve their models, so assume they may.
- We don’t train AI models on your content.
- AI output is made automatically and can be wrong. We don’t use it to make legal or similarly significant decisions about you.
What we use your information for
Privacy law in some places (such as the EU’s GDPR) asks us to name a “legal basis” for each use. We’ve put those in brackets.
- To run the Service for you (our contract with you): accounts, Teams, Agents, Computers, connections and email codes.
- To bill you (our contract, and our legal duties): subscriptions, credits, receipts and tax records.
- To keep Fungi secure and stop abuse (our legitimate interests): session IP addresses and browser details, logs, and enforcing the Acceptable Use Policy.
- To tell you about the Service (our contract, and our legitimate interests): security, billing and policy notices. We don’t send marketing email to product users unless you opt in.
- To count visits to fungi.computer (your consent): optional website analytics, only after you choose Accept.
- To obey the law (our legal duties): answering lawful requests and reporting child sexual abuse material.
Who has access to information about you
- Companies that provide parts of Fungi. Our Subprocessor list explains who they are, what they receive and which terms apply. See “What happens when AI works on your content?” for model providers’ data-use limits and uncertainties.
- Your Team. Team members can see the Team’s content, including Agent sessions and Computers in that Team, depending on their role.
- Services you connect, when an Agent acts through that connection.
- Authorities, when the law requires it, or to protect people from serious harm.
- A buyer or successor, if the business is sold or merged. They must keep the promises in this policy.
We don’t sell your personal information, and we don’t “share” it for cross-context advertising, as California law uses those words.
Cookies
- The signed-in product (my.fungi.computer) uses only the strictly necessary cookies that keep you signed in. They’re secure, and scripts on the page can’t read them. There’s no analytics or advertising tracking in the product.
- The public website (fungi.computer) asks before it uses optional analytics
through Google Analytics and, when configured, PostHog’s
eu.i.posthog.comendpoint. Optional analytics does not load until you choose Accept. We can remember either Accept or Decline in local storage. This is not a promise that the website or chat stores no other functional state. See “The launch list” below.
How long we keep your information
| Information | How long we keep it |
|---|---|
| Account, Team, Agent sessions, workspaces, Apps | Saved to run the service. There is no verified automatic expiry covering all these records. Deletion requests need store-by-store handling; closing a Team is not erasure. |
| Session IP address and browser details | Recorded with sign-ins. We have not verified an automatic purge deadline after a session ends. |
| Log of Team membership changes | We have not verified a fixed automatic deletion period. |
| Invitations | An invitation expiring prevents use; it does not establish deletion of its record or email address. No automatic 30-day purge is verified. |
| Billing and tax records | Some records may need to remain to meet legal duties. Ask which records and retention periods apply to your request; this policy does not establish one universal seven-year period. |
| Stripe records | Stripe has its own retention rules. Deleting a customer record does not establish that historical invoices, payment records or their personal fields were erased. |
| Backups | Copies are separate from live records. We have not verified a maximum age for every backup; deleting live data does not automatically erase backup copies. |
| Logs | Retention depends on the log store and receiver. There is no verified single 30-day limit for all logs. |
| Launch list | See “The launch list” below. |
Your rights, and how to use them
Depending on where you live, you may have the right to:
- see the information we hold about you
- correct it
- delete it
- get a copy to take elsewhere
- object to, or limit, some of the ways we use it
- withdraw your consent
Using these rights won’t change how we treat you.
How it works today. Fungi doesn’t have self-serve delete or export yet. To ask for anything, email privacy@fungi.computer from the address on your account.
- We’ll check it’s you, usually by replying to your account email.
- Deleting your data. Your information is stored in several places, so a deletion request needs individual review. We have not yet verified that our process removes every copy. Closing a Team, stopping a Computer or deleting one record is not enough. Ask us what can be deleted, what must remain and why, and how we will confirm the result. Backups, stored App files and records required by law need separate checks.
- If you own a Team with other members, we’ll ask you to hand it over to someone else, or close it, first.
- Getting a copy. Ask which data can be provided and in what format. We do not have a verified complete automated export of every store.
- Correcting your data. Use the available account settings, or email us for help. Never send us your sign-in codes, tokens or credentials.
- How quickly we’ll reply. Within 30 days.
- You can also complain to your local data protection authority (see “How to complain”).
The steps we take to keep your information safe
- Your information is encrypted while it travels between you and us. GitHub sign-in token storage is configured for encryption. This does not mean every file, log or user-supplied credential is encrypted at rest.
- Computers run in isolated virtual machines. Isolation does not make allowed files secret from an Agent or an authorized operator. Credentials copied into a Computer or conversation can remain there.
- No system is perfectly secure. If a breach affects your personal information, we’ll tell you, and the regulators, as the law requires.
Where your information goes
We’re based in the United States. The Fungi service uses operator-managed servers and the providers below. We have not independently verified one storage region for every service, backup and copy, so this policy is not a guarantee that all your data stays in the EU. Ask us for the regions relevant to your request.
- United States and China. Your prompts, the files your Agents read and the model’s answers go, through Cloudflare AI Gateway, to OpenAI in the United States and, for Kimi K3, to Moonshot AI in China, as described in “What happens when AI works on your content?”
- Video models. Video prompts go, through Cloudflare AI Gateway, to Google, ByteDance or MiniMax. We haven’t confirmed which countries they process them in.
- Cloudflare’s global network. Every request to Fungi passes through Cloudflare’s network, so it can pass through the country nearest to you. Cloudflare also handles App source code and the emails we send. The configured off-site backup job can upload encrypted copies to R2; its presence does not establish that a backup has run or that all backups are encrypted.
- United States and others. Stripe handles payments.
- PostHog. If you switch on analytics on the public website, your browser
sends events to the configured
eu.i.posthog.comendpoint. We have not independently verified the processing or storage region of every copy.
If you’re in the EU or UK, the law asks for a safeguard when your information leaves. We rely on the Standard Contractual Clauses in each subprocessor’s data processing terms where they provide them. See our Subprocessor list for which ones do.
The launch list
This section is about the list on fungi.computer, where people asked to hear when Fungi opens.
- If you joined the launch list, you’ll get one email when fungi.computer is ready. Joining doesn’t create an account or sign you up for a newsletter.
- Signup records, including your email address and signup time, are stored in Cloudflare Workers KV. Delivery and unsubscribe state also uses Durable Objects. Earlier sign-ups may include a name.
- We don’t send a confirmation email.
- We have not confirmed that signup records are automatically deleted within 30 days of the launch email. Removing a record used to send email does not necessarily remove your signup record.
- To be taken off the list, email privacy@fungi.computer, or use the unsubscribe link in the launch email.
- Optional analytics. The website asks first, with Accept and Decline side
by side. Until you choose Accept, it doesn’t load either analytics provider,
send either anything, or set cookies or local storage for analytics. If your
browser sends Global Privacy Control or Do Not Track, analytics stays off and
we don’t ask. If you accept, PostHog’s
eu.i.posthog.comendpoint receives the pages you visit on fungi.computer and a note that you signed up. We don’t send it your email address, what you typed into forms, the query part of the web address, or the page you came from, and we don’t build a profile of you. PostHog receives your IP address when your browser connects. We have not confirmed whether it keeps that address. PostHog keeps a random browser identifier in your browser’s local storage. We remember your choice, Accept or Decline, in local storage too. To change it, use “Privacy choices” at the foot of any page. Choosing Decline stops anything new being sent and clears PostHog’s storage from your browser. It doesn’t erase what was already sent; email privacy@fungi.computer and we’ll ask PostHog to delete it. - Google Analytics. The same Accept choice allows Google Analytics to count page visits. We send a fixed page address without query strings or fragments, a fixed page title and no referrer. We do not send form values or email addresses. Advertising signals and personalization are disabled. Google receives your IP address and browser information when your browser connects, and stores browser identifiers in first-party cookies. Decline stops collection and clears those cookies. It does not erase earlier events. Google’s processing is not restricted to the EU by this website.
- Holder perk. If you apply for the holder perk, we keep your email address and the wallet address you signed with, and use them only to contact you about that perk. Email support@fungi.computer to have them erased.
- Our fonts come from our own website. Your browser doesn’t contact Google Fonts.
Children
Fungi isn’t for anyone under 18. We don’t knowingly collect information from anyone under 18. If you think a child has an account, email us and we’ll delete it.
Who’s responsible for your information
Jamstack Consulting LLC is responsible for keeping your information safe. If you’re worried about the information we hold, contact privacy@fungi.computer.
How to complain
We take complaints seriously. If you’re unhappy with how we’ve handled your information, email privacy@fungi.computer. If you’re the letter-writing type, send your envelope to Jamstack Consulting LLC, 13005 West Flower Street, Avondale, Arizona 85392, United States.
You can also complain to the data protection authority where you live.
Changes to the policy
When we change this policy, we’ll post the new version here with a new version number and date. If a change is significant, we’ll email account holders at least 30 days before it takes effect.