FAQ
Answers to the questions founders ask before they start.
Honest answers to the questions we get most often. Including the sharpest critiques from people who don't think the Product Brain category is real yet.
- What is ProductLobster?
- ProductLobster is the agentic Product Brain — context engineering for product teams. It holds your product's compounding memory and runs the work a senior PM team would run, for founders and operators building products without a dedicated PM. It's the PM-vertical implementation of the Context Engineering discipline Gartner named in 2025.
- Who is ProductLobster for?
- Solo founders, indie hackers, technical founders, designers, marketers, and operators building products without a dedicated PM and without the budget to hire one. People already using AI tools to build (Lovable, V0, Cursor) AND to think (ChatGPT, Claude). Teams of 1 to 5 with real product stakes. The champion is the founder. The purchase is direct.
- How long does the first audit take?
- Five minutes to a live workspace with PM chat already grounded in your product. Focused analyses (competitive landscape, customer research, prototype) run about 5 minutes each. The full audit (5 stages, one strategic checkpoint) takes about 60 minutes. The full launch analysis (7 stages, three strategic checkpoints) takes about 60 to 90 minutes.
- Why not ChatGPT or Claude with project memory?
- ChatGPT and Claude both have persistent memory and projects where you upload files. Both give you context that survives between sessions. Neither structures that context for product work. ProductLobster's Product Brain is a knowledge graph typed by problem, customer, competitor, opportunity, decision, hypothesis, experiment, roadmap, and outcome — each with a citation to the source it came from and a timestamp showing how recent it is. So an analysis can ask "which decisions are stale," "which competitors shipped what," or "where is the evidence thin" and get a real answer, not a paragraph that summarizes a transcript. The two compose: ProductLobster can read your chat tool's project files; your chat tool will be able to read the Product Brain via MCP (coming soon). Memory is the substrate. PM-workflow ontology is the irreducible difference.
- Why not Claude Skills plus a CLAUDE.md for product work?
- Claude Skills are hundreds of tokens of procedural knowledge each, loaded on demand when the model decides a task needs them. They teach Claude how to do something. They don't store your product's state across sessions — that lives in CLAUDE.md or MEMORY.md, which is a file an engineer maintains for engineering work. ProductLobster is the layer your Skills should read from. The Product Brain is queryable, decision-grade state: customer evidence with provenance, competitive landscape over time, locked strategic frames, prototype iterations, checkpoint decisions. Skills plus CLAUDE.md is a great pattern for code work, and we use it ourselves. For product work, you want the procedural knowledge to run over a product-specific state layer that compounds week over week. That state layer is what ProductLobster ships. The two compose: if you use Claude Skills today, ProductLobster's MCP server (coming soon) is what they read from when the question is about your product.
- How is this different from Sentra, Hyperspell, or Glean?
- Horizontal Company Brains capture whole-org memory through a generic ontology. They answer questions like "what did Sarah promise the customer?" The Product Brain captures the product's memory through a product-specific ontology. It answers questions like "should we build feature X or feature Y given everything we know about our customers, competitors, and roadmap?" Horizontal brains will eventually add product use cases as features. ProductLobster will eventually integrate with them as sources. Neither replaces the other. The product-specific ontology is the part you can't get elsewhere.
- Why not Letta, Mem0, Zep, or Cognee?
- Those are memory infrastructure for developers. They expose APIs for engineers to build agent applications. ProductLobster is workflow software for product teams that uses a Product Brain as the durable internal asset. Different buyer. Different shape of value. ProductLobster may integrate with them under the hood. We don't compete with them at the buyer layer.
- I'm not a PM. Will I understand the output?
- Yes. The methodology powers the thinking invisibly. The output is plain language. PM frameworks like JTBD, switching forces, segmentation, strategy kernel, and shaping are applied without requiring PM vocabulary. You describe your product in your own words; the system runs senior PM work over your product's memory and shows you the result you can act on.
- Why subscription instead of per-run?
- Subscription is the core model because the Product Brain compounds week over week — pure-per-purchase pricing would punish the use case that makes ProductLobster valuable. The Free tier gives you one full audit or launch to feel the activation moment; Plus removes the lifetime ceiling and adds monthly recurring depth. Hybrid overages on top cover the depth-on-demand month: when you need one more audit before a launch decision, or one more prototype to show a customer, buy it on-demand in human-readable product units. No tokens. No opaque limits.
- Why do you have overages? Isn't that the PAYG thing you said you removed?
- Two different things. What we removed: a separate Pay-As-You-Go tier sold INSTEAD of a subscription. That confuses buyers and cannibalizes the recurring product. What we added: hybrid overages — on-demand units added to your existing subscription, in human-readable product units (one more audit, one more prototype). The mental model is "I'm a Plus subscriber who bought one extra LAUNCH this month," not "I'm a PAYG customer." Every AI tool at scale runs hybrid (Cursor, Lovable, GitHub Copilot post-2026). The discipline we hold: overages are NEVER in tokens, NEVER framed as a percentage of a limit, NEVER opaque. Always concrete units you can predict.
- What if Anthropic ships Claude PM?
- Auto Memory has already shipped. What's specifically ProductLobster: the Product ontology, the opinionated workflow, the methodology, and the Brain commands (queries that read your product's state and actions that run PM work from inside any agent). Anthropic can ship better memory. They're unlikely to ship the opinionated PM workflow that's 15 years of product-leadership distilled into a product. And if they do, ProductLobster becomes the workflow any agent (including Claude PM) can read from via MCP.
- Isn't this just rebranded enterprise search?
- Enterprise search is read-mostly retrieval over existing data. The Product Brain is write-discipline-aware structured memory with decision-grade state changes. We don't replicate Glean's strength (broad enterprise data ingestion). We serve a different shape of value: PM-workflow-native context with opinionated methodology. Different buyer, different job. Glean serves the enterprise PM org at scale. ProductLobster serves the founder and small team that horizontal incumbents don't economically reach.
- Is this a real category, or a Twitter bubble?
- Gartner named context engineering a core enterprise discipline in 2025. Adobe hires Context Engineers at $215K. AWS selected Mem0 as exclusive memory provider for its Agent SDK. Y Combinator named Company Brain as one of 15 official Summer 2026 RFS categories. The category has crossed three of four signal streams for real formation. The fourth (mainstream PM-podcast coverage) is thin, and is exactly what our content pillar is designed to close. Roberta Micore is right about the category as it's being sold to 15-person teams by enterprise vendors. Her critique is the exact reason our beachhead is the founder who can't afford to hire a PM, not the enterprise procurement team. Different size of tool for a different size of farm.
- Security and compliance?
- Workspace-isolated per user. Three-layer IP defense: prompt-level instructions, output leakage detection, and KG leak prevention. Pipeline circuit breakers. Token budgets. Content Security Policy enforcement for prototype rendering. Clerk auth (SOC 2 Type II provider). Enterprise SSO and audit logs are Phase 3 only. If you need them today, ProductLobster is not yet the right fit.
- Migration or switching cost?
- None. Onboarding is the first analytical step. Workspace creation takes about five minutes. Document upload is drag-and-drop. No data migration from prior tools is required. The Product Brain compounds from the moment you start.
- Is the Product Brain my data, or yours?
- Your data. Workspace-isolated. You own everything in your Product Brain. Full export available anytime.
- Can I bring my own LLM?
- Not at MVP. ProductLobster ships with OpenRouter-backed model selection. Bring-your-own-key is on the roadmap. Not in v7 Now.
- Does ProductLobster integrate with Linear, Jira, or Notion?
- Read-side: Granola transcript ingestion, Gmail thread monitoring, weekly briefing, and proactive monitoring are Phase 1 agentic capabilities and are listed as Coming Soon on the roadmap. We don't pin a specific quarter — initial scope only. Write-side (creating Linear tickets from analyses) is not currently on the roadmap. We claim what's shipped and what we're building toward; speculative "Later" features stay off the site.
- What happens to my Product Brain if I cancel?
- Read-only access for 30 days. Full data export available before account closure. We don't delete your data unless you ask.
- Is this for me if I only want a PRD?
- Probably not. ChatPRD or Disko is faster for the discrete PRD-generation job. ProductLobster's value is the compounding Product Brain. If you're not running product work over time, the subscription is over-tooled for the job.