Answers, plainly put.

PGS AI is an experimental AI cognition workspace built around multi-core parallel cognition. Where most AI runs a single model, PGS AI's multi-core builds run several specialized cognitive cores in parallel and synthesize them into one response.

It's a research-grade platform shipped as a consumer product, currently in paid beta.

No. PGS AI runs in your web browser at ai.pgsgrove.com. No installation, no app store. Just sign in and start using it.

Any modern web browser on desktop, laptop, tablet, or phone. We recommend Chrome, Safari, Firefox, or Edge in their current versions for the best experience.

Visit ai.pgsgrove.com to create an account. You'll choose a tier, set up your account, and be in the workspace within a few minutes.

No. We don't offer a free tier. Our entry tier (Seedling, $4/mo) is priced as low as we can make it while still keeping the platform private and high-quality for everyone.

Free tiers in this industry are typically used to gather data on users. We chose a different path: pay a small amount, keep your privacy, get genuine quality. The full reasoning is on our main page.

PGS AI is restricted to adults (18 or older, or the age of majority in your jurisdiction, whichever is greater). The Service is also not currently available in some regions.

A build is a specific cognitive architecture: which cores are active, how they coordinate, and what the synthesizer trunk does with their output. PGS AI ships in three lines (Sunbird, Kephra, Phoenix), and each line has both a single-core and a multi-core variant.

For a deep dive on each build, see our architecture page.

Multi-core builds run several specialized cognitive processors in parallel. One core handles emotional reasoning, another handles spatial and dimensional thinking, another handles strategic analysis, and so on. An executive synthesis layer (the trunk) weaves their outputs together into one response.

You can see each core's contribution in the multi-core thinking panel that comes with every multi-core response.

Three lines, each with its own cognitive style.

Sunbird is fast and expressive, built for quick conversation and casual exploration.

Kephra is the power-user's daily driver, with native chain-of-thought reasoning for serious creative and technical work.

Phoenix is the deepest-thinking line, built for long-horizon tasks, deep analytical synthesis, and multi-step reasoning.

Each line ships as both a single-core build and a multi-core build (Sunbird3C, Kephra4C, Phoenix4C).

Single-core is faster and direct. Use it when you want a quick, focused answer or you're iterating rapidly.

Multi-core is deeper and produces more dimensional responses. Use it when you want emotional, spatial, and strategic perspectives blended into the answer, or when you want to see the cores at work in the multi-core thinking panel.

Yes. You can change builds at any point in a conversation. The new build will pick up the conversation context and respond from there. Memory is shared across builds within the same Node.

For every multi-core response, a thinking panel surfaces what each core contributed: how the emotional core read your message, how the strategy core analyzed the path forward, how the pieces came together. Click to expand it on any response.

Yes. New builds typically arrive monthly. Higher tiers get early access to experimental builds. We're also building a seven-core build currently in research.

Seedling and Daily access Sunbird and Kephra. Plus and above unlock the full lineup including Phoenix and all multi-core builds. Power and Ultra get early access to experimental builds. Full breakdown on the pricing section of our main page.

Memory is layered. Conversation history, canvas drafts, AI-saved notes, knowledge core uploads, and in-chat uploads all become part of a multi-dimensional memory the AI can draw from naturally. You don't start over between sessions.

Full breakdown on our memory page.

A Node is an isolated AI workspace with its own memory, conversations, and personality. Build one for work, one for creative projects, one for personal use. Nothing leaks between them. Each Node remembers only its own world.

From the main interface, use the Node switcher to create or switch between Nodes. Each one is independent from the moment you create it.

The number of Nodes scales with your tier. Higher tiers unlock more. The architecture scales as your work does.

Yes. Saved memories are visible from your settings. You can review every note the AI has saved, edit anything that's wrong, and delete anything you want gone. No hidden memory.

Yes. Individual memories, full conversations, or your entire history. Whatever scale you want. Your memory belongs to you.

The Knowledge Core is where you can upload reference materials, project documents, and other content you want the AI to draw from. Anything you put there is available to the AI in any conversation in that Node.

Tap the voice icon and start speaking. Your AI responds in voice. Tap to interrupt anytime. The build you're using stays the same in voice mode: same cores, same memory, same intelligence. We don't swap you to a smaller, faster model when you switch to voice.

A growing selection of high-quality voices, each with its own character. Pick whichever fits your AI configuration. New voices arrive regularly.

Yes. Tap to interrupt at any time. The AI receives the interruption and responds to where the conversation actually went, not where it was planning to go.

Canvas is a live editable workspace for documents, code, and longer-form content you're building together with the AI. Anything you create in Canvas stays accessible in your Node memory and can be referenced later.

Yes. Canvas includes a code execution environment for supported languages. You can write code, run it, and iterate together with your AI.

No. Your conversation content is never used to train AI models, ours or anyone else's. This is a deliberate architectural choice and a core privacy commitment of the Service. The full details are in our Privacy Policy.

Your conversations are stored in association with your account so the Service can function as intended. Conversations are not shared with third parties for marketing, not sold, and not used for training.

Authorized human reviewers may access conversations in specific safety circumstances, described below.

Authorized reviewers may access conversation content when our automated safety systems flag content that needs human review, when you submit a report through the in-app "Report a Concern" function, when you contact support requesting help that requires us to look at your conversation, or when we have a good-faith belief that review is necessary to investigate suspected violations or prevent imminent harm.

This is a meaningful disclosure and we want users to know about it. Full details in our Privacy Policy.

No. We do not sell personal information. We do not share personal information for cross-context advertising. We do not work with data brokers.

Content that sexualizes or harms minors, operational instructions for serious harm, sexual content, hate and harassment, malicious code, and a few other categories. Our full Community Guidelines cover the specifics.

Use the "Report a Concern" button inside the app. Reports go to a real human review queue and shape how the Service improves over time. You can also email support@pgsgroveinternal.com.

PGS AI runs on infrastructure in the United States and other regions where our service providers operate. Conversation content is processed transiently by our AI infrastructure partners during inference and is not retained on disk by them in the ordinary course of operations. Account and conversation data is stored in our database with appropriate access controls.

Tier breakdowns are on the pricing section of our main page. Briefly: Seedling ($4) is for casual exploration with Sunbird; Daily ($12) unlocks Kephra; Plus ($25) unlocks Phoenix and all multi-core builds; Pro ($50) and Power ($99) scale up usage and add experimental access; Ultra ($199) is day-zero access to everything we ship.

Yes. Change tiers anytime from your account settings. Changes take effect at the start of your next billing period.

Cancel anytime from your account settings. You'll keep access through the end of the current billing period.

Subscription fees are generally non-refundable. When you cancel, you keep access through the end of your current billing period and aren't charged further.

Refunds may be granted at our discretion in exceptional circumstances. Email support@pgsgroveinternal.com if you have a specific situation to discuss.

You'll receive a notice and have time to update your payment method. If the issue isn't resolved, your account may be paused until billing is current.

The AI industry standard is to offer free tiers as a way to gather data on users. That puts overhead pressure on paid tiers and lowers quality for everyone.

We've instead opted to give our users true privacy from the moment they log in, and keep quality high for everyone. Our entry tiers are priced to be accessible, so that everyone can have privacy and quality without exception.

PGS AI is currently offered as individual subscriptions. Phoenix Grove Systems also offers B2B consulting and custom AI system builds for organizations. Contact us at support@pgsgroveinternal.com if you're interested in either.

Major credit and debit cards through our payment processor. Specific accepted methods may vary by region.

PGS AI is in active development. The platform is fully usable today, but you'll see new builds, new features, and improvements arriving regularly, often monthly. Things will evolve. Some experimental builds become permanent. Some get refined into something better.

You're not paying for a finished product frozen in time. You're paying for a platform that keeps growing alongside you.

New builds and meaningful features typically arrive monthly. Smaller improvements and fixes arrive throughout each month.

Builds may be improved over time, but the way they think and the lineup names stay stable. We'll communicate meaningful changes through the app and email. Your Nodes, conversations, and memory remain yours through any changes to underlying builds.

Beta software has bugs. If something breaks, please report it through the in-app "Report a Concern" function or email support@pgsgroveinternal.com. We track every report and prioritize fixes accordingly.

Email us at support@pgsgroveinternal.com with the subject line "Feature suggestion." We read every one. We can't promise we'll build everything, but user input genuinely shapes what we work on next.

Currently in research. We'll announce when it's ready for consumer release. Higher tiers will likely get early access first.

From the sign-in page at ai.pgsgrove.com, click "Forgot password" and follow the steps. You'll receive a reset link at the email address on your account.

From your account settings inside the app. If you've lost access to the email on file, contact support.

Try a hard refresh first (Cmd-Shift-R on Mac, Ctrl-Shift-R on Windows). If that doesn't work, clear your browser cache for ai.pgsgrove.com. If issues persist, email support@pgsgroveinternal.com with details about what you're seeing and which browser you're using.

Most voice issues are browser permissions related. Make sure your browser has microphone access for ai.pgsgrove.com. Also check that no other app is currently using your microphone exclusively.

You can request account closure from your account settings or by emailing support@pgsgroveinternal.com from the address on file. Conversation content is deleted following the wind-down period described in our Privacy Policy, except where retention is required by law.

Email support@pgsgroveinternal.com. We aim to respond within one to two business days. For specific issues, please include your account email, what you were trying to do, and any error messages you saw.

Our Terms of Service and Privacy Policy are both linked here. The Section 2.3 of the Terms covers PGS AI specifically. Our Community Guidelines describe what is and isn't permitted on the Service in plain language.