Terms, privacy
& community rules.

A1. What this draft does—and does not do

SeshFriends is the selected product name; the earlier prototype was called Weed Spot. This document describes current pilot practices and proposes a launch EULA, privacy framework, and community rules. The proposed terms are not being activated as a new binding launch contract. No one has consented to future AI screening or data sales by acknowledging this draft.

Before launch, identify the legal operator, publish working support and privacy contact channels, review the agreement for each supported jurisdiction, implement the required choices and recordkeeping, and replace this draft with a dated final version. The service is not currently a verified legal-compliance or child-safety system.

A2. Information used in the pilot

The original ChatGPT-hosted pilot uses ChatGPT sign-in. The independent SeshFriends deployment uses Supabase email/password accounts, email verification, recovery emails delivered through Resend, and secure session cookies. Both supply an authenticated account identifier and email to recognize members and the owner. Supabase processes passwords; the application database does not store them. Independent beta access is limited to invited email addresses. Short-lived login rate-limit records contain hashed IP/email identifiers. It stores limited security-review access logs and your chosen display name, profile design, optional profile details, visibility choices, buddy preferences, home type, timezone, Smoke Circle connections, messages, task completions, earned-Bud transactions, posts, likes, comments, activity read status, reports, submitted repair guides and their review decisions, and files you upload. In-app reminder hours and delivery-slot history are saved locally in this browser for each account; closed-app push delivery is not active. It does not store your ChatGPT password.

Camera access starts only after you choose a capture action and grant browser permission. Capturing does not itself upload or publish a file. Photo uploads are re-encoded in the browser, but identifying details can remain visible in the picture. New videos are optimized to low resolution before upload and may contain location or other metadata. Review the material before sharing.

The Munchie Map concept page has been removed from this beta; live business discovery is not available. The app does not currently request continuous GPS tracking or maintain a history of dispensary visits. Hosting and sign-in providers may process connection and security information under their own terms. Verify their logging and retention practices before launch.

A3. Visibility, messages, and your controls

Private and unsearchable are separate default settings. You can make your profile public within the pilot without making it searchable, or allow search while keeping detailed profile content Circle-only. A direct link to a private profile exposes a limited name/avatar card so another member can request a connection. Acceptance is required for private details and messaging. Your name or avatar can also appear in requests and interactions.

An uploaded draft and a shared post are different. Drafts and task-history records stay private to the account. Each new post or video can be private, Circle-only, or explicitly Public to all signed-in SeshFriends members. A Public post is visible to members even when its author’s profile is private. Existing posts that follow profile privacy keep that audience until changed. Profile details and cover photos follow profile visibility. Profile pictures can appear in member interactions. Search can find permitted shared posts by caption and author display name, and searchable members or accepted friends by display name. Private drafts and messages are not included in search. Sharing a post link does not change its audience; recipients must be allowed to view the original, which may later become unavailable. Likes and comments are visible to members who can currently view the post, subject to blocking and moderation. The post owner receives in-app activity notifications with your display name and avatar. You can remove your own comment; post owners can also remove comments on their posts. Reading activity alerts does not mark private messages read. A public post is not confidential. Other people may capture or copy content they were allowed to view; changing visibility cannot recall those copies.

The owner has no general app screen or API permission to browse unshared uploads or unrelated conversations. Selected reported content has a limited, logged review path. These are application access controls, not end-to-end encryption or a cryptographic guarantee against someone controlling infrastructure or deployment. Direct messages are deliberately text-only. Private image messaging is deferred; Circle-only feed posts are the available way to share photos with accepted friends. The app must not claim operator-inaccessible private image messaging until a suitable encrypted design and independent security review support that claim.

Private messages are accessible through the app only to the participants in an accepted connection. They are not end-to-end encrypted. A participant report permits the owner to open only that selected message for a stated moderation purpose. Evidence stays hidden in the ordinary report list until an explicit review action, and that action is logged. Removing or blocking the connection stops new messaging and access to the thread through the app. These access controls are not a promise that a hosting provider cannot process stored data.

Account deletion removes primary account records and uploaded files associated with the account, including sent and received messages and Circle connections. Limited security-access logs and reports submitted by other members can remain; their retention periods must be finalized before launch. Deleting a post removes its attachment when no remaining app record uses it. Abandoned uploads can remain until account deletion. Copies outside the app, provider logs, and backups may have different lifetimes; no universal deletion deadline is promised in this draft. Establish retention and lawful-preservation procedures before launch.

A4. Current safety operations

The pilot supports user reports, blocking, owner review, content removal, and account bans. It does not currently run AI sales screening, image/video moderation, government-ID verification, or an automated reporting pipeline to law enforcement. Automated detection should never be represented as active until the implementation, disclosure, and any required consent are in place.

The current pilot requires an adult age attestation, not identity or age verification. Under-21 participation is prohibited even where a younger patient might otherwise have medical-cannabis access. The operator must establish a child-safety response process before broader use. Do not upload suspected child sexual abuse material to an ordinary support channel or redistribute it to other users.

1. Eligibility and account responsibilities

For the proposed service, you must be at least 21, have legal capacity to agree, and use the service only where your participation and the relevant features are lawful. You must not impersonate others, evade a suspension, lend an account to a minor, or misrepresent information used for eligibility. Cannabis laws and medical/recreational access differ by location; an account does not establish eligibility to buy or use cannabis.

You are responsible for protecting your sign-in and for activity you authorize. Report suspected compromise using the published support channel. The operator may restrict features, geography, accounts, or access when reasonably needed for safety or lawfulness. A self-attestation alone is not proof of verified age.

2. License to use the service and ownership

The operator grants a personal, limited, revocable, nonexclusive, nontransferable license to access the service for permitted use. The operator or its licensors retain rights in the app software, original branding, interface, and original assets. You must not bypass security controls, scrape private accounts, interfere with infrastructure, or exploit another member.

You retain ownership of content you create. You grant the operator a nonexclusive, royalty-free license to host, copy, transmit, format, display according to your visibility choices, and moderate that content as needed to operate the service, including through necessary processors. This is not a blanket transfer of your content or an unrestricted license to sell private communications. Personal-information rights remain governed by applicable law and the privacy notice.

3. No cannabis sales or transaction arrangements

No member may offer, solicit, broker, arrange, or complete cannabis or cannabis-product transactions through the service, including through posts, photos, video captions, profile text, comments, or direct messages. This rule applies even if a transaction might be legal elsewhere. “DM for prices,” coded sale offers, personal menus, delivery/pickup arrangements, payment instructions, and off-app handoffs intended to arrange a cannabis transaction are prohibited.

The current beta has no live location finder. A future version may help users find licensed dispensary and restaurant locations and learn about cannabis culture. A location listing is not a license endorsement, a sales channel, a reservation, or a guarantee that a business may lawfully serve a user. No cannabis checkout or peer-sales feature is authorized by these terms. Confirmed cannabis-sales activity is grounds for immediate account termination and withdrawal of shared content, subject to a meaningful appeal process.

4. Community rules and child protection

No racism, discriminatory harassment, bullying, threats, stalking, doxxing, fraud, sexual exploitation, grooming, trafficking, child sexual abuse material, nonconsensual intimate material, or other illegal conduct. Do not seek out minors, encourage minors to use cannabis, or use cannabis offers to exploit someone. Do not post private information, faces, recordings, or copyrighted work without the necessary rights or consent.

The service is not an emergency response channel. Use the appropriate local emergency or child-protection authority when immediate intervention is needed. User-facing reporting should identify urgent child-safety concerns, restrict access to reported material, and route it to trained, authorized reviewers. A prohibited act can justify removal or termination regardless of the account’s Bud balance.

5. Buds, customization, and a free service

The core service is intended to remain free. Buds are earned app credits for cosmetic items, not money, cryptocurrency, an investment, transferable property, or a means of buying cannabis. They cannot be cashed out or used to pay another member. All offered decorations should remain earnable through ordinary participation, without paid-only exclusives or pressure to purchase.

Purchased Buds are not offered in the pilot. Any future optional support or purchased-credit program requires separate pricing, payment, refund and consumer-rights terms before activation. It must not be introduced by implying that existing users already consented to it. No current contract should promise refund mechanics or ownership of purchased balances that do not yet exist.

6. Moderation, appeals, and reporting

The operator may investigate reports, restrict distribution, remove content, suspend features, or terminate accounts for confirmed violations. Safety restrictions may be applied promptly while a review is pending. Provide notice of a decision and a way to appeal when lawful and appropriate; do not publish private reporter details. Document evidence, reviewer access, the reason for a decision, and any correction.

Reports must be made in good faith. A report about a private message authorizes review of that selected message and the information the reporter submits, not unrestricted browsing of unrelated conversations. Repeated or coordinated reports alone should not determine guilt. AI output, if later used, can be wrong; uncertain or contested cases need an appropriate review path.

7. Proposed automated screening—clear notice required

CURRENT STATUS: AI screening is not active. PROPOSED FEATURE: before delivering posts or messages, automated systems may analyze text for cannabis sale offers, coded solicitations, “message for price” language, attempts to arrange transactions, threats, fraud, exploitation, or other serious violations. The purpose is safety enforcement, not permission to sell the contents of messages.

Lawful consensual adult nudity is not by itself evidence of illegal conduct. If private media is later supported, normal staff access should be prohibited, private content must not be sold or used for advertising profiles, and any exception must identify a specific reported item or a validated serious-safety flag, an authorized role, purpose, limited duration and an audit trail. An AI flag is not proof of a crime. Human review of intimate material must be minimized; child-safety duties and lawful process still apply. A server-side scanner can process the media it scans, so do not promise that this is simultaneously an operator-inaccessible, end-to-end encrypted channel without a design that actually supports the stated threat model.

Before activating private-message screening, show a prominent notice in signup and messaging, identify the categories analyzed, explain any external processing provider and retention, state how automated decisions affect delivery or accounts, and obtain and record any legally required consent. Do not hide this disclosure in a long agreement or treat a buried clause as sufficient notice. If required screening is a condition of messaging, explain that plainly before the user chooses the feature.

Photo, video, and audio screening would require separately described capabilities. Do not imply that scanning text detects every sale, every abusive image, or every threat. Messages would remain not end-to-end encrypted unless the architecture is changed and independently verified. A new or materially different use requires updated notice and any legally required choice before it starts.

8. Data commercialization—no current data sales

The pilot does not sell personal information to data brokers. The operator’s interest in a future data business is not a present authorization. Neither registering, accepting this draft, nor granting a content license gives the operator unrestricted ownership of users’ personal information or waives nonwaivable privacy rights.

Any future commercial sale or sharing program must first identify the actual data categories, purposes, recipients, jurisdictions, and applicable legal restrictions; publish a specific notice; and implement all required consent, opt-out, access, correction, deletion and appeal mechanisms before transfers occur. Respect recognized universal opt-out signals where required. Core access must not be conditioned on surrendering rights that applicable law protects.

Cannabis-related inferences can expose sensitive circumstances. Private-message contents, precise location, medical information, intimate media, credentials, minors’ data, and sensitive inferences require particular scrutiny and may be subject to restrictions that make a proposed sale impermissible. This draft does not authorize transfers of those categories. Private media and message contents must not be included in a broker-sale or advertiser-access program. Do not treat advertising services, aggregated reporting, sale of personal information, and transfer of identifiable media as interchangeable.

9. Lawful cooperation and protection of children

The operator may respond to valid legal process, make legally required reports, preserve records when required, and make emergency disclosures only as applicable law permits. Validate the requesting authority, legal basis, scope, and urgency. This clause does not create a standing police backdoor, remove legal-process requirements, or authorize indiscriminate sharing of private messages.

Where provider-reporting duties apply to apparent child sexual exploitation, the operator will follow the applicable reporting and preservation requirements, including reporting to the National Center for Missing & Exploited Children when required. Access must be limited to authorized personnel and lawful recipients. The launch plan needs a documented child-safety contact, evidence-handling process, reviewer training, and escalation procedure.

Notify affected users about legal requests when legally permitted and appropriate, subject to restrictions such as court orders and legitimate child-safety or emergency concerns. The operator must not promise advance notice in every case or use an informal request as a substitute for the legal authority required to disclose communications.

10. Buddy behavior, guides, and third-party information

The buddy’s positive/roast personality and profanity setting do not change community rules or authorize harassment. Repair walkthroughs, generated explanations, and business information can be incomplete or wrong. Use appropriate judgment, follow safety stopping conditions, and get qualified assistance for hazards. If you are impaired, choose safe containment or defer the repair; do not use the app as justification to drive or perform dangerous work.

The service does not provide emergency, medical, legal, or professional repair services. Listings, maps, external links, and future sponsors have their own terms. Clearly label paid placements. A sponsored listing does not grant access to users’ private tasks or messages. No third-party endorsement should be implied without authorization.

11. Changes, termination, and deletion

Users may stop using the service and use available account-deletion controls. The operator may suspend or terminate accounts for confirmed prohibited conduct or legal obligations, subject to applicable rights and an appropriate appeal path. Retention must be tied to a documented operational or legal purpose; legal preservation and necessary security records can affect deletion where permitted or required.

Material changes to privacy practices, screening, paid features, or commercial data sharing must be explained before they begin. Obtain a new choice or consent where required; do not retroactively claim that this draft approved a new use. Version notices and acceptance records should identify exactly what the user was shown.

12. General legal provisions to finalize

To the extent permitted by applicable law, the service is provided as available without a guarantee of uninterrupted operation, complete accuracy, or absolute security. No provision excludes liability or remedies that cannot lawfully be excluded. Any liability limitation, governing-law selection, venue provision, or dispute process must be finalized for the operator and jurisdictions before launch.

No mandatory arbitration or class-action waiver is inserted in this draft. A court’s finding that one clause is unenforceable should not invalidate the remaining lawful provisions. A final agreement must identify the contracting operator and how users can contact support, exercise privacy rights, report safety issues, and appeal moderation.

Operator checklist before adoption

1. Confirm legal operator, business address, product name/trademark clearance, and support, privacy, child-safety and legal-request contacts.
2. Have counsel review state/federal consumer, communications-monitoring, cannabis-discovery, age/access and privacy obligations; decide supported jurisdictions.
3. Finalize current-provider disclosures, security controls, retention/deletion schedules, backup behavior and lawful preservation exceptions. The current deletion feature is not a legal-hold system.
4. Publish a final EULA, privacy policy, community rules and reporting procedure. Add clear, versioned acceptance and any separate required consents; do not log consent to inactive or undisclosed features.
5. Before AI moderation: connect a model securely, identify provider handling, test false positives/evasion and child-safety escalation, restrict reviewer access, and add prominent messaging notice and required consent.
6. Before private-media messaging: choose an audited encryption and key-management architecture, separate operator/infrastructure roles, prevent general staff browsing, define narrowly scoped safety/report exceptions, and test deletion, access logs and legal-request handling. Do not describe access controls as a guarantee against infrastructure administrators.
7. Before any data-sale program: determine whether it is lawful, implement all applicable rights and opt-out/consent workflows, document vendors and transfers, and assess sensitive-data restrictions. No broker export is enabled by this draft.
8. Complete real-device and multi-account testing; review age assurance and child-safety operations before admitting a wider audience.

Research references for legal review

These sources inform the draft; they do not establish that every cited law applies or that the app is compliant.

• FTC: deceptive designs, buried terms and privacy choices
• California Attorney General: privacy rights and sensitive information (applicability must be assessed)
• 18 U.S.C. § 2702: restrictions and exceptions for disclosure of stored communications
• 18 U.S.C. § 2258A: provider reporting and preservation duties
• FTC: COPPA guidance, including general-audience services with actual knowledge
• FTC: enforcement involving sales of sensitive location data