Privacy Policy
Last updated October 13, 2026
1. Who this policy covers, and our two roles · 2. What we collect · 3. How we use it · 4. Who receives it · 5. Cookies and browser storage · 6. Where it is kept · 7. How long we keep it · 8. Security · 9. Your choices and rights · 10. Changes to this policy · 11. Contact
Pidgen AI (“Pidgen”, “we”) operates Pidgen, a hosted service that plans software work and runs coding agents on machines our customers control, and this website, pidgen.ai. This policy says what personal information we handle, why, who receives it, how long we keep it, and what choices you have.
1. Who this policy covers, and our two roles
- When you visit this website, ask for access, sign in, or administer an Account, we decide how your personal information is used. We are the controller of that information.
- When your organisation puts work into Pidgen (Tickets, comments, agent transcripts, source code that appears in them), your organisation decides what goes in and why. Your organisation is the controller of that content and we process it on its behalf, under our Terms of Service. If you have a question about content your organisation placed in Pidgen, ask your organisation first.
Pidgen is a service for businesses. It is not directed at anyone under 18, and we do not knowingly collect personal information from children. During the beta, Pidgen is offered only in the United States.
2. What we collect
Information you give us
| When | What |
|---|---|
| You request access | Your name, work email, company, team size, the forge you use, and any message you add. |
| You accept an invitation or sign in | Your email, display name, and password. We store the password only as a one-way hash. |
| You turn on two-factor sign-in | A TOTP secret, stored encrypted, and recovery codes, stored only as hashes. |
| You create an API token | A label and a hash of the token. We show the token once and cannot show it again. |
| You contact us | What you send us. |
Information created when you use the service
| What | Detail |
|---|---|
| Sessions | When each session started and was last used, a device label taken from your browser’s description of itself, and a shortened form of your IP address. We do not store your full IP address with the session. |
| Your actions | An audit record of administrative actions in an Account (for example invitations, role changes, token and worker changes), with who did them and when. |
| Usage | How much each Account uses the service: runs, their duration, and token counts that workers report. |
| Setup progress | The time each Account first reached each setup step. |
| Operational logs | Requests to the service and this website, including IP addresses, and errors, kept to run and secure them. We redact secrets from logs. |
Work content your organisation puts in
Tickets, Intents, comments, steering messages, uploaded files, workflow definitions, and what workers report about each run: agent transcripts, artifacts, change snapshots, and pull-request and branch facts. A transcript can contain source code and anything else an agent read on the worker’s machine. Work content can include personal information if your organisation puts it there.
Information about workers
A worker runs on a machine your organisation controls. It tells us a host identifier, its version, the versions of required tools and coding-agent programs, whether each coding agent is logged in, and whether the machine’s git credentials reach each repository. Paths in what it reports can reveal the machine’s hostname and the operating-system username.
Credentials your organisation gives us
A Forge Connection’s access token, and, if your Account opts in to decision nodes with its own key, that key. We encrypt these with a key held for your Account, never show them again after they are saved, and use each only for your Account.
What we do not collect
- Model-provider credentials. They stay on your workers. Pidgen never receives them.
- Files on your workers, beyond what is described above.
- Payment details. The beta takes no payment.
- Advertising or analytics trackers. We use none, on the service or on this website.
3. How we use it
| Purpose | Information used |
|---|---|
| Deciding on access requests and creating Accounts | Access-request information |
| Signing you in and keeping your Account secure | Account, session and two-factor information |
| Providing the service: planning work, assigning it to your workers, showing results, opening pull requests with your forge token | Work content, worker information, credentials |
| Sending service email: sign-in links, password resets, invitations, access-request outcomes, and notices that an Account is suspended, locked or closing | Email address |
| Measuring use against an Account’s limits | Usage |
| Supporting you and fixing problems | Whatever is needed for the request, including work content when you ask us to look |
| Understanding where setup fails, so we can improve it | Setup progress |
| Securing the service and this website, and meeting legal obligations | Logs, audit records |
We do not sell personal information. We do not share it for advertising. We do not train machine-learning models on your content, and content from one Account is never used for another Account.
4. Who receives it
-
People in your Account. Members see work content and worker information according to their roles. Owners and Administrators see the Account’s audit log. Whoever administers a worker’s machine can read the source code and transcripts of every Ticket that worker runs.
-
Our subprocessors, the companies that process information for us:
Subprocessor What it does for us What it processes Applies to Cloudflare, Inc. Carries traffic between your browser and our servers, and protects them IP addresses and the content of requests and responses in transit Everyone who uses the service or this website MXroute Delivers service email Recipient email addresses and the content of service email Every Account TypeSafe Decision provider for decision nodes The input of each decision node, which can include prompts and Ticket content Only Accounts that opt in to decision nodes We will update this list before a new subprocessor starts processing customer information, and tell Account Owners by email.
-
Your forge. When Pidgen opens or merges a pull request, it acts on your forge with the token your organisation supplied.
-
Your model providers. Your workers send prompts and code to the model providers you configured, under your own accounts. That traffic does not pass through Pidgen, and those providers are not our subprocessors.
-
Authorities, when the law requires it. We will tell the affected Account first unless the law forbids it.
-
A successor, if our business is sold or merged. This policy continues to apply to information already collected.
Access by Pidgen staff
A small number of authorised Pidgen administrators can open an Account’s data to operate and support the service. During the beta this access is not recorded in your Account’s audit log. Before general availability, every such access will be logged, attributed, time-limited, and visible to your Account’s Owners.
5. Cookies and browser storage
This website sets no cookies. The service uses only what it needs to work:
| Name | Kind | Purpose |
|---|---|---|
pidgen_session |
Cookie | Keeps you signed in. |
pidgen_csrf |
Cookie | Protects your session against forged requests. |
| Preferences | Browser storage | Your theme, the Account and Project you last used, and list and navigation settings. They stay in your browser. |
We set no advertising cookie and no third-party cookie.
6. Where it is kept
The service and this website run on servers we operate in the United States, reached through Cloudflare’s network. If you use Pidgen from another country, your information is transferred to and processed in the United States.
7. How long we keep it
| Information | Kept |
|---|---|
| Access requests | For the length of the beta, or until you ask us to delete yours. |
| Your profile, credentials and sessions | Until your user is erased. |
| Work content, events and transcripts | For the life of the Account, during the beta. |
| Account audit log | For the life of the Account. |
| A closing Account | Work stops for a 30-day window in which closing can be cancelled. Then the Account’s live data is removed and a complete export is kept in an isolated archive for 90 days, unless a legal hold applies. A minimal record that the Account existed and was closed remains. |
| Backups | 3 days. |
| Files on your workers | Not ours to keep or delete. Workspaces, repository clones and agent transcripts on machines you control stay there until you remove them. |
Erasing a user. When a user is erased we delete their profile, credentials, sessions, tokens and memberships. Records of past events refer to people only by an internal identifier, which then shows as “Deleted user”. Content the user wrote inside an Account belongs to that Account and stays with it.
8. Security
We encrypt information in transit, encrypt customer-supplied credentials with per-Account keys, keep each Account’s data separate from every other’s, offer two-factor sign-in, and keep backups. No system is perfectly secure. If we learn that a security incident has led to unauthorised access to, or loss or disclosure of, an Account’s data, we will tell that Account’s Owners by email without undue delay and no later than 72 hours after we confirm it.
9. Your choices and rights
- See and correct your profile in the app.
- End other sessions and revoke tokens in the app.
- Ask us to erase your user, or for a copy of your information, at [email protected]. We answer within 30 days. If the request concerns content in your organisation’s Account, we will refer you to your organisation.
- Depending on the state you live in, you may have further rights under its law. Write to us and we will tell you how we apply them.
We do not discriminate against anyone for using these rights. Service email is necessary to use the service and cannot be turned off. We send no marketing email.
10. Changes to this policy
We will post a new version here and change the date at the top. If a change is material, we will tell Account Owners by email before it takes effect.
11. Contact
Pidgen AI · 50 Sunbeam Acres Lane, Sedona, AZ 86351 · [email protected]