Privacy Policy
Welcome to Signal Flow. This website is a development, testing, and demonstration environment used solely to showcase technical features and test functionality, including user authentication via Google Sign-In.
Because this is a testing environment, please do not use real personal data, sensitive information, or production credentials when interacting with this site.
Who runs this site
Signal Flow is run by Higherwave (higherwave.org) as a private, invitation-only demonstration of a research-to-requirements tool. It is not a commercial service, and it is shared only with people who have been invited. Questions about this policy can be sent to david.garrett@higherwave.org.
What Signal Flow is, and why it handles information
Signal Flow lets a team plan a research study, upload interview transcripts, review the themes found in them, and turn those themes into requirements and stories. To keep the demonstration limited to invited people, signing in is done with a Google account.
Information we receive from Google when you sign in
When you choose "Continue with Google", Google shares these basic profile details with Signal Flow:
- your name,
- your email address (and whether Google has verified it), and
- your profile picture address, if you have one.
Signal Flow requests only the basic sign-in scopes (openid, email, and profile). It does not request, read, or store anything from your Gmail, Google Drive, Calendar, Contacts, or any other Google service, and it cannot act on your Google account.
How we use that information
- To check that your email address has been invited, and to give you the access level (owner, editor, or viewer) that was set for it.
- To show your name and role in the app, and to record who made a change.
- To keep you signed in for the length of your session.
We do not use this information for advertising, marketing, profiling, or analytics, and we do not sell it or give it to anyone for those purposes.
Information you enter in the app
Anything you type or upload, such as project names, participant details, and transcript text, is stored so the app can show it back to you. This is a testing environment: please use made-up or sample data only.
Prototype thumbnails
When an editor links a prototype in the Design step, our server opens that link once, on our own systems, to take a small picture of the page for the table. The picture is stored with the requirement and the link is not sent to anyone else. A page that needs a sign-in shows the sign-in page, so use the upload option for private prototypes. Private network addresses are never visited. An owner can switch this off.
AI features
Signal Flow can use an AI service, Anthropic (anthropic.com) through its API, for three jobs. Each is switched on or off separately from the rest of the app, and each is started only by an invited editor or owner, never by a visitor using the read-only demo.
- Analysis. The text of the transcripts in the project being analyzed, each labeled as a session with a customer or a conversation with your own team, together with the study's journey stages, the project description, the questions to ask and tasks to observe from the task list, and the names, short definitions, and alternative names of the experience categories your team keeps, is sent so the model can find themes and say which experience category each is about.
- Drafting requirements. To write each step of a requirements document, the app sends the study's description and expected outcomes, the roles of the participants, the title, summary, and suggested opportunity of each theme included in the document, one short quote from a participant for each of those themes, and the steps already written in the same document.
- Matching themes. To avoid listing the same problem twice across projects, the app sends the title, a short description, and the suggested opportunity of each new theme, together with the titles of the themes already shared across projects, and the name of the experience category each belongs to when it has one. The same details are sent when the app compares shared themes to find ones that say the same thing. No transcript text is sent for this.
Before anything is sent, the app removes email addresses and phone numbers and, for transcripts, replaces speaker names with labels such as P1 and removes participant names it recognizes; file names are not sent. Your name, email address, and profile picture from Google are never sent to an AI service. Per Anthropic's API terms, Anthropic keeps this input and the model's output for 30 days by default for safety and security purposes. Signal Flow has not joined any Anthropic program that shares data for model improvement. When AI is switched off, or its monthly budget or credit is used up, the app uses a simulated keyword version of these steps and nothing is sent to an AI service. Because this is a testing environment, please enter only made-up or sample data.
Slack
When a Slack connection is set up and a team has chosen a channel for its launches, the app posts to that channel in Higherwave’s Slack workspace (Slack, slack.com): the launch’s name, team, target date and owner’s name with a link to it, and then, in a thread under that message, short notes when answers are asked for, when a step is answered (the step, the answer, who gave it and the reason they wrote) and when the final decision is made. Anyone in that channel can read these. Nothing is read back from Slack, and approvals are given only in the app. Without a Slack connection, or for a team with no channel, nothing is sent. Because this is a testing environment, please enter only made-up or sample data.
Answering from Slack
A person can connect their Slack account to their account here with a one-time code, so that a button on a launch’s Slack thread records their answer as them. Slack sends the app the Slack user’s id, the workspace id and what the person typed in the answer form (the reason, or a link and its title), signed so the app can tell it came from Slack; the app keeps the Slack user id with the connection and the answer or link in the launch record. The app does not read channel messages: it sees a message only when a person chooses “Add link as evidence” on it, and then only the link they confirm. Because this is a testing environment, please enter only made-up or sample data.
Following a launch
An editor can add colleagues to follow a launch and share an early digest with them. The digest is built from the project’s own data without any AI rewriting: the themes the research found, a few customer quotes as they were said (never with the speaker’s name), and the plain-language brief. It is sent by email to the followers and, if the team has a Slack channel, posted in the launch’s Slack thread, where everyone in that channel can read it. A follower gets the digest and brief only, not access to the research. Because this is a testing environment, please enter only made-up or sample data.
Jira tickets
When an editor chooses Create Jira ticket and a Jira connection is set up, the app sends what is shown in that dialog (the summary, description, issue type, priority, labels and, if chosen, the assignee) to Higherwave’s Jira site (Atlassian, atlassian.com), where it becomes an issue. Nothing is sent until an editor confirms the dialog, and the description is the requirement as written in the app, with a link back to it. Without a Jira connection the ticket is only recorded here and nothing is sent. Because this is a testing environment, please enter only made-up or sample data.
Cookies
Signal Flow sets one first-party cookie, a session cookie (named like better-auth.session_token), that keeps you signed in for up to seven days or until you sign out. It is needed for sign-in to work. We do not use advertising or analytics cookies.
Where information is stored, and who else handles it
The information above is stored in a Postgres database hosted by Neon (in the United States). The site is hosted by Vercel, which may keep ordinary server logs, such as IP addresses and requests, as part of hosting. Google provides the sign-in service, and Anthropic handles transcript text and research summaries when the AI features are on (see above). These providers handle information only to run the service for us. We do not share your information with anyone else, except if the law requires it.
How long we keep it, and how to have it removed
Because this is a testing environment, we keep information only while the environment exists, and it may be reset or deleted at any time without notice. An owner can remove your access on the Members page, which also ends your session. To have your name, email address, and sign-in record deleted, email david.garrett@higherwave.org and we will remove them. You can also withdraw Signal Flow's access to your Google account at any time at myaccount.google.com/permissions.
Google API Services User Data Policy
Signal Flow's use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements.
Children
Signal Flow is not directed at children, and it is not for people under 18.
Changes to this policy
If this policy changes, for example if the app starts using another outside service, this page will be updated and the date above will change.