SuperApp brings team conversations, AI models and shared work into one place

SuperApp is a collaborative AI workspace where a team’s conversation, source files and finished work can stay together. It combines shared threads with access to multiple AI providers and the creation of documents, presentations and other artifacts. The company was previously Instabase: its renamed corporate identity now spans this collaborative product and the continuing AI Hub enterprise document offer.

01 / ProductOne current company identity, two distinct product contexts

The official rename announcement, dated 17 August 2026, says Instabase Inc. is becoming SuperApp Inc. It explicitly states that AI Hub continues and existing customers remain supported. This blueprint therefore uses SuperApp as the company identity instead of presenting Instabase and SuperApp as separate companies.

The SuperApp product page describes real-time shared threads, multi-provider AI and collaborative creation of documents, slides, spreadsheets and interactive artifacts. Its central idea is that people can shape work with AI in the same conversation where the sources and feedback live. That is a coordination proposition as much as a model-access proposition.

AI Hub Automate remains a distinct enterprise document workflow product. Its documented scope includes extraction, cross-document validation, human case review and governed release workflows. A SuperApp team subscription should not be assumed to include AI Hub production processing. Conversely, continuing support for an AI Hub deployment does not establish that every new SuperApp collaboration feature is part of that enterprise agreement.

02 / AudienceFor teams whose AI work loses context between tools

A small project team preparing a research-backed recommendation is a plausible SuperApp audience. One person gathers evidence, another questions assumptions and a third owns the final presentation. If the AI conversation is private to the first person, later reviewers may see a polished answer without the sources or decisions that shaped it. A shared workspace can make that reasoning process easier to inspect.

The product is less compelling when the actual problem is simply choosing a single model for occasional private questions. Collaboration introduces its own work: deciding who belongs in the thread, which files are authoritative and when a draft becomes final. A team should evaluate whether shared context reduces handoffs enough to justify changing an established working pattern.

The Notion blueprint offers a comparison for teams organizing durable knowledge and documents. The Superhuman blueprint covers another productivity company and product context; Superhuman and SuperApp are separate identities despite the similar names. The useful comparison is where the team’s discussion, knowledge and deliverable should live, rather than assuming that access to many models alone resolves collaboration.

03 / WorkflowA proposed team research brief with a traceable review path

Consider a product team preparing a market-entry brief from approved public research and internal planning notes. This is a proposed SuperApp evaluation, not a tested workflow. Choose a low-sensitivity project with a clear final deliverable and invite only the people who need to participate. Establish the authoritative source files before asking AI to synthesize them.

Create one working thread for the project and assign a human owner for the deliverable. Ask the AI to separate supported facts from assumptions and unresolved questions. A shared conversation makes those distinctions visible to the reviewers, but it does not make them correct automatically. Each consequential claim should still be checked against its actual source.

Use the collaborative artifact as the evolving brief, while reviewers leave specific corrections in context. Test whether a late-arriving colleague can understand why an important recommendation changed. That is a useful evaluation of the product’s coordination value: the team should spend less time reconstructing decisions, not merely produce more revisions.

Where a paid plan permits manual model selection, compare a second model on one difficult question using the same evidence. Do not treat agreement between models as independent corroboration; both may repeat the same unsupported assumption. The exercise is valuable if it exposes a missing source, conflicting interpretation or poorly framed question for the team to resolve.

Before finalizing, ask the owner to identify the accepted version and preserve the decisions that matter. SuperApp’s enterprise page describes a ledger of inputs, model requests, tool activity and outputs. That enterprise capability is not proof that every self-serve tier exposes the same record. Demonstrate the actual revision and activity access available in the selected plan.

Export the agreed deliverable and check whether a colleague outside the pilot can use it without losing essential references. Also remove one collaborator in a controlled test and verify the resulting access boundary. Success is a coherent brief with understandable revisions and audience control; it is not simply a faster first draft or the largest number of available model names.

04 / PricingCredits, seats and enterprise scope need separate treatment

The current product pricing is in US dollars excluding applicable taxes. The table records selected monthly offers relevant to a collaboration evaluation. Credits are a usage allowance, not a promise of a fixed number of prompts or completed projects. Model routing, media generation and the complexity of work can change the consumption pattern.

Annual payment options are also displayed, and should be read as annual commitments rather than cancellable monthly prices. For example, Team Pro is shown at $250 per member billed yearly; the lower monthly equivalent is a presentation of that yearly charge. Confirm the selected billing period and current checkout details when purchasing because the public page says plans and prices may change.

The Terms of Service permit personal or internal business use, require users to be adults and describe automatic renewal. They also reserve model availability and apply restrictions to unauthorized automated access and service-bureau resale. The proposed pilot uses ordinary internal collaboration, not an assumed backend API entitlement. Enterprise deployment controls and AI Hub commercial arrangements need their own confirmed scope.

05 / DistinctionsShared context is the product’s most useful hypothesis

SuperApp’s distinctive proposition is that team feedback and AI activity can remain attached to the work being created. That can reduce a familiar failure mode: a document reaches a reviewer with the latest wording but without the questions, discarded assumptions or source changes that explain it. The proposed pilot deliberately measures whether another person can reconstruct those decisions.

The enterprise material adds a broader deployment proposition, including dedicated infrastructure, model choice and activity records. These are supplier-described capabilities. No portability or failover test was conducted for this blueprint, and illustrated ledger examples are not customer results. An enterprise evaluation should demonstrate how a real task behaves when a model or connected tool is unavailable.

AI Hub’s history makes the company relevant to buyers already familiar with enterprise document processing, but the newer collaborative workspace has a different job. AI Hub Automate emphasizes configured document applications, validation and controlled promotion. SuperApp emphasizes people and AI creating work together. Understanding that separation prevents a buyer from transferring assumptions about one product’s deployment and support to the other.

06 / QuestionsConfirm visibility, provider handling and export completeness

Who can see the thread and its files after an external collaborator joins? The terms explain that workspace administrators can control access and that shared material may be copied or redistributed by recipients. Before a real rollout, test membership changes and sharing with the type of information the team actually plans to use.

How is content handled when requests use multiple providers? The terms say content is not used to train the general-purpose models offered through the service, while relevant information can be transmitted to model providers for the requested work. The Services Privacy Policy describes service-provider processing, customer instructions and international transfers. These statements should not be simplified into a claim that information never leaves one environment.

What survives export? The enterprise page describes exportable conversations, files, artifacts and agent state, but the practical format, self-serve plan access and import into another system need demonstration. A downloadable final document may be sufficient for one team; another may need the revision record and source attachments. Define that requirement before treating portability as established.

For existing AI Hub customers, clarify the supported path for expansion and the product-specific contract. The rename notice explicitly commits to continued support, but it is not a public universal tariff or a guarantee that every deployment option is offered to every new buyer. Keep the support commitment and the commercial question distinct.

07 / DecisionTest whether the team can finish and explain the work together

SuperApp is worth evaluating when a team’s AI-assisted work is slowed by fragmented discussions, files and revisions. Use a bounded project and measure how easily participants can reach the source evidence, understand changes and identify the accepted deliverable. That is a more revealing test than comparing a few impressive model responses.

Select the plan for its actual collaboration, usage and record-access requirements. Keep AI Hub evaluation separate where the job is enterprise document automation. The company’s changed identity is clear, but the right product choice still begins with the team’s specific work and the controls it needs to complete it.