The most useful AI assistant is the one that remembers everything about you. The problem is, that memory is currently owned by the assistant, not by you. It’s a fragmented, proprietary asset locked inside each model or application you use. Bot Food, a Toronto-based startup, is building a tool it calls RaLHF to change that. The product is a personal context layer, a private and portable library for a user’s digital life that any AI can query with the user’s permission [ralhf.ai/]. It’s an ambitious bet on data ownership as the next wedge in consumer AI.
The Wedge: Portable, User-Controlled Context
RaLHF, pronounced "Ralph," operates as a background service. It signs into a user’s applications and files, catalogs the information it finds, and stores it in an encrypted digital library [BetaKit, October 2026]. Users can view, edit, and selectively share this context with different AI systems. The company describes it as "infrastructure for personal digital memory" that belongs to the individual and works across platforms [ralhf.ai/about]. The core proposition is efficiency and control: instead of every new AI chatbot starting from scratch or building its own incomplete profile, it can tap into a single, user-maintained source of truth. For a user, it means an AI could know your flight preferences, dietary restrictions, and project history without you having to repeat yourself each time.
The Team Behind the Bet
The founders are betting their experience in scaling and selling enterprise software can translate to a consumer-facing data product. CEO Ian Collins is a repeat founder with seven startups under his belt and a track record of raising over $100 million, including co-founding Wysdom.AI, which was sold in 2024 [ralhf.ai/about]. President and COO Nitin Singhal was COO of Wysdom.AI through that acquisition and previously helped scale telecom software firm Redknee from 50 to 5,000 employees [ralhf.ai/about]. The technical leadership is similarly seasoned. CTO Brendan Stennett co-founded and built the open-data platform ThinkData Works, and VP of AI and Engineering Faik Faisal led development of Amazon Q Developer [ralhf.ai/about]. This is a team built to handle complex data pipelines and enterprise-grade engineering, now focused on an individual user’s problem.
The company’s board includes Jeff Brunet, former CTO of Wysdom.AI and now managing partner at GreenSky Ventures, which led Bot Food’s pre-seed round [BetaKit, October 2026].
| Role | Name | Key Background |
|---|---|---|
| CEO & Founder | Ian Collins | Founded 7 startups; co-founded Wysdom.AI (acquired 2024). |
| President & COO | Nitin Singhal | COO of Wysdom.AI; former CTO/VP R&D at Redknee (Optiva). |
| CTO | Brendan Stennett | Co-founder & CTO of ThinkData Works. |
| VP of AI & Engineering | Faik Faisal | Led Amazon Q Developer; former roles at Slack, BlackBerry. |
| Board Member | Jeff Brunet | Former Wysdom.AI CTO; Managing Partner, GreenSky Ventures. |
Early Traction and Funding
Bot Food is in its earliest stages, operating with a reported team of seven people [BetaKit, October 2026]. The company has raised a total of $3 million CAD, anchored by a pre-seed round led by GreenSky Ventures with participation from Extreme Venture Partners [BetaKit, October 2026]. The company launched its first product, Mémoire, in October 2025, and has since released public Claude Code plugins for RaLHF and related tools, indicating an active development and community-engagement strategy [news10.com, October 2025] [awesomeclaudeplugins.com, retrieved 2026].
The Technical Breakdown
Architecturally, RaLHF sits between the user and their AI applications. It must solve several hard problems concurrently: secure credential management for dozens of potential service integrations, efficient and incremental data synchronization, a query interface that respects user-defined sharing rules, and local encryption that still allows for performant retrieval. The decision to build public plugins for platforms like Claude is a smart distribution tactic, putting the tool directly in the workflow of early adopters. The real technical moat, however, won't be the connectors but the proprietary algorithms for structuring, deduplicating, and making this heterogeneous personal data instantly useful to an LLM. That’s a non-trivial data engineering challenge.
The Scale and Skepticism Test
The bet is compelling, but the path to scale is lined with significant hurdles. The product asks consumers to undertake a new, ongoing maintenance chore,managing a data library,for a benefit that is currently theoretical for most. The value proposition is strongest for power users who already juggle multiple AI tools. Furthermore, the company must navigate intense privacy scrutiny; a breach in a service designed to hold a user’s entire digital footprint would be catastrophic. There’s also a strategic risk from the very platforms RaLHF aims to complement. If OpenAI, Anthropic, or Google decide to build their own cross-platform, user-portable memory layers,and they have the data to start with,it could instantly commoditize the need for a third-party tool. Bot Food’s answer likely hinges on neutrality and trust, positioning itself as the Switzerland of personal context, but that’s a difficult position to own against integrated giants.
For now, Bot Food has assembled a credible team, secured early capital, and identified a genuine friction point in the evolving AI stack. The next twelve months will be about proving that friction is painful enough for users to adopt a new infrastructure layer. Key milestones to watch will be the publication of an integration roadmap, any measurable user growth metrics, and whether the model evolves to include a clear path to monetization. The founders have built and sold a company before; this time, they’re betting on the individual user’s right to own the data that makes AI personal.
Sources
- [ralhf.ai] RaLHF - The context layer for AI | https://ralhf.ai/
- [ralhf.ai] About RaLHF | https://ralhf.ai/about
- [BetaKit, October 2026] Bot Food secures $1.5 million to give AI users more control over their data | https://betakit.com/bot-food-secures-1-5-million-to-give-ai-users-more-control-over-their-data/
- [news10.com, October 2025] Bot Food Corporation Introduces Mémoire, A User-Owned AI Context Platform Protecting Personal Digital Data | https://www.news10.com/business/press-releases/ein-presswire/855800265/bot-food-corporation-introduces-memoire-a-user-owned-ai-context-platform-protecting-personal-digital-data/
- [awesomeclaudeplugins.com, retrieved 2026] botfoodai/RaLHF-plugins | Awesome Claude Plugins | https://awesomeclaudeplugins.com/botfoodai/RaLHF-plugins