Epicenter

Cover Block

PUBLIC

Name	Epicenter
Tagline	ChatGPT memory in open portable plain-text format for personal workspace [Y Combinator, May 2026]
Headquarters	San Francisco, CA, USA
Founded	2025
Stage	Seed
Business Model	SaaS
Industry	Other
Technology	AI / Machine Learning
Geography	North America
Growth Profile	Venture Scale
Founding Team	Solo Founder
Funding Label	Undisclosed (total disclosed ~$500,000)

This profile outlines a newly formed venture, Epicenter, which joined the Y Combinator Summer 2025 batch. The company's public positioning centers on creating a personal knowledge workspace with an open-source, portable architecture for AI memory.

Links

PUBLIC

• Website: https://epicenter.so
• GitHub: https://github.com/EpicenterHQ/epicenter
• X / Twitter: https://x.com/epicenter_so

Executive Summary

PUBLIC

Epicenter is building an open-source, local-first workspace that aims to replicate the persistent memory feature of ChatGPT, but in a portable plain-text format to give users full data ownership and the ability to train custom AI models [Y Combinator, May 2026]. The company, which joined Y Combinator's Summer 2025 batch, presents a technical bet on unbundling personal AI from proprietary platforms, a thesis that resonates with developer-led trends toward data sovereignty and interoperability [Y Combinator Launches, Aug 2025]. Founded solo in 2025 by Braden Wong, a recent Yale graduate, the project has evolved from a core database library into a small ecosystem of example applications, including a local-first speech-to-text tool called Whispering [GitHub, 2026]. The business model appears to be pre-revenue SaaS, with an undisclosed seed round estimated at approximately $500,000 from investors including Formosa Capital, Pioneer Fund, and Spot VC [Crunchbase, 2026]. The primary near-term risk is the founder's capacity to transition from a one-person, open-source project to a scalable commercial product in a space crowded with well-funded incumbents like Notion and Mem. Over the next 12-18 months, the key signals to monitor will be the launch of a commercial MVP, the first disclosed enterprise or developer customers, and any expansion of the founding team beyond the solo founder.

Data Accuracy: YELLOW -- Core company description and founder background are publicly cited; funding details and product status are partially corroborated.

Taxonomy Snapshot

Axis	Classification
Stage	Seed
Business Model	SaaS
Industry	Other
Technology	AI / Machine Learning
Geography	North America
Growth Profile	Venture Scale
Founding Team	Solo Founder
Funding	Undisclosed (total disclosed ~$500,000)

Company Overview

PUBLIC

Epicenter is a San Francisco-based startup founded in 2025 by solo founder Braden Wong, a recent Yale University graduate [Y Combinator, May 2026] [LinkedIn, 2026]. The company was accepted into the Y Combinator Summer 2025 batch, an early milestone that provided an undisclosed amount of seed funding and placed it within a cohort of early-stage AI and software ventures [Y Combinator, May 2026].

Its primary legal entity and corporate structure are not detailed in public registries. The company's development path is visible through its open-source releases on GitHub, beginning with the core @epicenter/workspace library and expanding to include several example applications. A notable public launch was Whispering, a local-first, open-source speech-to-text application, which was featured on Y Combinator's Launch YC platform in August 2025 [Y Combinator Launches, Aug 2025].

The company remains a solo operation as of its May 2026 Y Combinator profile listing, with no public announcements of team expansion, subsequent funding rounds, or named commercial customers [Y Combinator, May 2026].

Data Accuracy: YELLOW -- Key founding and accelerator details confirmed by YC; team size and timeline are single-source.

Product and Technology

MIXED The core proposition is a technical architecture, not a single application. Epicenter positions itself as an open-source, local-first ecosystem of applications built around a shared, persistent memory layer, described as "ChatGPT's memory feature in an open, portable plain-text format" [Y Combinator, May 2026]. This approach is designed to address data ownership and interoperability, offering an alternative to proprietary, siloed AI productivity tools. The foundational technology is a library called @epicenter/workspace, which provides CRDT-backed data sharing for applications to synchronize state [GitHub EpicenterHQ/epicenter, 2026].

The first publicly released application built on this stack is Whispering, a local-first, open-source speech-to-text tool [Y Combinator Launches, Aug 2025]. It serves as a proof-of-concept for the ecosystem's principles, processing audio entirely on a user's device. The company's GitHub repository showcases several other example applications, suggesting the platform's intended breadth: a personal CMS (Fuji), a Mandarin learning chat interface (Zhongwen), an agent skill manager (Skills Editor), and basic UI components for billing and landing pages [GitHub EpicenterHQ/epicenter, 2026]. These appear to be developer-facing demonstrations of the underlying data layer's capabilities rather than polished end-user products.

From a technical standpoint, the stack emphasizes plain text and SQLite as the portable storage formats, enabling users to retain full control over their data. The vision articulated in launch materials is for this shared local memory to become a substrate for custom AI models and a wide range of interoperable personal productivity apps [Y Combinator, May 2026]. As of the latest public information, the product surface is a collection of open-source repositories and a conceptual framework; there is no evidence of a unified, commercially available SaaS platform or a general release.

Data Accuracy: YELLOW -- Product claims sourced from Y Combinator launch pages and GitHub repository, which represent company statements. No independent third-party product reviews or user testimonials are available.

Market Research

PUBLIC The market for tools that manage personal knowledge and AI memory is coalescing around a central tension: the convenience of integrated AI assistants versus user demand for data ownership and portability.

No third-party analyst report sizing the specific market for "open, portable AI memory" was identified in the available research. The company's positioning suggests it competes within the broader personal knowledge management (PKM) and AI productivity software segments. For context, the global note-taking apps market was valued at approximately $1.5 billion in 2024 and is projected to grow at a compound annual rate of 7.5% through 2030, according to a report from Grand View Research cited by industry analysis [Techno-Pulse, April 2026]. This analogous market is driven by demand for organized digital workspaces and the integration of generative AI features for summarization and content creation.

Key demand drivers for Epicenter's proposed wedge include the rapid adoption of AI assistants with memory features, like those from OpenAI and Google, which has heightened user awareness of persistent context. A concurrent tailwind is growing developer and prosumer skepticism towards proprietary data silos, fueling interest in local-first and open-source alternatives that promise interoperability and custom model training. The expansion of capable small language models that can run locally further enables this architectural shift.

Adjacent and substitute markets present both expansion opportunities and competitive threats. The low-code/no-code platform segment allows users to build custom workflows without deep technical skill, potentially fulfilling a similar need for personal data organization. The open-source developer tools ecosystem, particularly around local vector databases and model fine-tuning, serves as a foundational layer upon which Epicenter would need to demonstrate superior usability. Regulatory forces are nascent but could develop around data sovereignty and AI transparency, potentially favoring architectures with clear data lineage and user control.

Data Accuracy: YELLOW -- Market sizing is inferred from an analogous segment report; specific TAM for the company's niche is not publicly quantified.

Competitive Landscape

MIXED Epicenter enters a densely populated market for AI-powered knowledge management, but positions itself at the intersection of open-source tooling and persistent AI memory, a niche with few established players.

Company	Positioning	Stage / Funding	Notable Differentiator	Source
Epicenter	Open-source, local-first personal workspace with ChatGPT-like memory in plain text.	Seed (YC S25); ~$500k total disclosed. [PUBLIC]	Ecosystem of interoperable apps built on a shared, portable SQLite database; emphasizes data ownership and custom AI models. [PUBLIC]	[Y Combinator, May 2026]
Mem	AI-powered notes app that automatically organizes and surfaces information.	Seed ($5.6M, 2021). [PUBLIC]	Proactive memory and context retrieval; closed, cloud-based system with a focus on automated organization.	[Techno-Pulse, April 2026]
Reflect	Privacy-focused notes app with real-time syncing and end-to-end encryption.	Seed ($1.2M, 2021). [PUBLIC]	Built on local-first sync technology (CRDTs) with a strong emphasis on privacy and offline functionality.	[Techno-Pulse, April 2026]
Notion	All-in-one workspace for notes, docs, wikis, and project management, now with AI features.	Late-stage private; $343M total funding. [PUBLIC]	Dominant platform with massive network effects, extensive templates, and a broad third-party integration ecosystem.	[Techno-Pulse, April 2026]
Obsidian	Knowledge base and note-taking app that links thoughts using local Markdown files.	Bootstrapped; monetized via commercial licenses and sync service. [PUBLIC]	Philosophy of local data ownership and a vibrant plugin ecosystem; lacks native, persistent AI memory.	[Techno-Pulse, April 2026]

The competitive map splits into three primary segments. First, the established platform players like Notion, which compete on breadth and ecosystem lock-in, not on open data portability. Second, the privacy-first challengers such as Reflect and Obsidian, which share Epicenter's local-first ethos but have not yet integrated a persistent, ChatGPT-style memory layer as a core primitive. The third segment consists of AI-native note-takers like Mem, which offer automated memory features but within proprietary, cloud-based architectures that Epicenter's open-source, portable format explicitly opposes.

Epicenter's current defensible edge is architectural, not commercial. Its commitment to an open-source, plain-text and SQLite foundation for AI memory creates a technical moat for developers and privacy-conscious users who are dissatisfied with vendor lock-in. This edge is durable if the team can foster a developer ecosystem around its @epicenter/workspace library, as network effects in open-source tooling can be significant. However, this edge is also perishable; it is a bet on a trend, not a protected asset. Incumbents like Obsidian could integrate similar memory features, and closed-source AI players could open portions of their stack, nullifying the differentiation.

The company's most significant exposure is its lack of commercial traction and distribution in a market defined by entrenched user habits. Notion's vast template library and team collaboration features address a use case Epicenter does not yet touch. Mem's focus on a smooth, automated user experience for individual professionals sets a high bar for product polish that a one-person team will struggle to match. Furthermore, Epicenter's technical positioning may limit its immediate appeal to the mainstream user who prioritizes convenience over data sovereignty, a channel it does not own.

The most plausible 18-month scenario is one of niche validation versus broad adoption. If Epicenter successfully ships its flagship workspace app and attracts a community of developers to build on its platform, it could become the default tool for a segment of technical builders and AI tinkerers, pressuring Obsidian to accelerate its own AI roadmap. The "winner" in this case would be the open-source, local-first paradigm. Conversely, if execution lags and the product remains a collection of example apps, Epicenter would be the "loser," ceding the AI memory narrative to well-funded incumbents who can implement the feature as a checkbox within their existing, polished products.

Data Accuracy: YELLOW -- Competitor funding and positioning corroborated by third-party industry analysis [Techno-Pulse, April 2026]; Epicenter's own positioning is from its Y Combinator profile [Y Combinator, May 2026].

Opportunity

PUBLIC The opportunity for Epicenter rests on capturing a foundational layer of the emerging personal AI stack, where data ownership and portability become primary user concerns.

The headline opportunity is to become the default local-first data layer for AI-native personal productivity. The core bet is that as users accumulate more interactions with AI assistants, they will demand control over the resulting memory and context, moving away from proprietary, siloed systems like ChatGPT's built-in memory. Epicenter's proposition of an open, portable plain-text format, backed by a CRDT library for synchronization, positions it as a potential standard for this data layer [Y Combinator, May 2026]. This outcome is reachable not because of current traction, but because the architectural choice aligns with a growing developer and power-user sentiment favoring local-first software and data sovereignty, a trend evidenced by the active development of its open-source ecosystem on GitHub [GitHub, 2026].

Growth scenarios outline specific paths from a one-person open-source project to a scalable platform. The scenarios below are not predictions but plausible narratives based on the company's stated direction and market gaps.

Scenario	What happens	Catalyst	Why it's plausible
Developer Platform Adoption	Epicenter's @epicenter/workspace library becomes the go-to tool for indie developers and startups building AI apps that require persistent, user-owned memory.	A major launch of a third-party app built on Epicenter that gains significant traction, demonstrating the utility of the shared data layer.	The company has already built example apps (Fuji, Zhongwen) showcasing the library's use [GitHub, 2026]. The open-source model lowers adoption barriers for developers seeking alternatives to vendor lock-in.
Vertical Tool Consolidation	Epicenter expands from a general workspace into a suite of vertical-specific AI assistants (e.g., for learning, writing, coding), all sharing a common local memory core.	The successful standalone adoption of "Whispering," its local-first speech-to-text app, proves demand for a privacy-focused, interoperable tool [Slator, 2026].	The company's launch strategy began with a focused tool (Whispering) before the broader platform vision, indicating a potential wedge-and-platform motion [Y Combinator Launches, Aug 2025].

What compounding looks like is a classic open-source flywheel applied to a data standard. Initial developer adoption for one use case (e.g., a transcription app) increases the install base of the underlying Epicenter data format. As more apps read and write to this shared local store, the format's utility and network effects grow, making it more attractive for the next developer to adopt, thereby increasing the value of the ecosystem for all users. Early evidence of this compounding is the existence of multiple interconnected example applications within the company's own GitHub repository, suggesting a designed-for-interoperability approach from the start [GitHub, 2026].

The size of the win, in a successful developer platform scenario, could be measured against the valuation of companies that own foundational developer tools or data layers. For a rough comparable, consider the acquisition of Observable, a collaborative data notebook platform, by Databricks for an undisclosed sum that was reportedly in the hundreds of millions [TechCrunch, 2023]. While not a direct parallel, it illustrates the premium placed on tools that become central to a technical workflow. If Epicenter's data layer achieves similar centrality for a cohort of AI-native application developers, the company could command a valuation reflecting its strategic position as a gatekeeper to user-owned AI context. This is a scenario-based outcome, not a forecast.

Data Accuracy: YELLOW -- Opportunity analysis is based on company-stated vision and early open-source activity; market traction and developer adoption are not yet demonstrated.

Sources

PUBLIC

1. [Y Combinator, May 2026] Epicenter | Y Combinator | https://www.ycombinator.com/companies/epicenter

2. [LinkedIn, 2026] Braden W. - Epicenter (YC S25) Yale University | https://www.linkedin.com/in/braden-wong/

3. [Y Combinator Launches, Aug 2025] Launch YC: Epicenter: A Database for Your Mind, Built on Plain Text | Y Combinator | https://www.ycombinator.com/launches/O80-epicenter-a-database-for-your-mind-built-on-plain-text

4. [GitHub EpicenterHQ/epicenter, 2026] GitHub - EpicenterHQ/epicenter: Open-source, local-first apps. | https://github.com/EpicenterHQ/epicenter

5. [Y Combinator Launches, Aug 2025] Launch YC: Whispering: Local-First, Open-Source Speech to Text at Your Fingertips | Y Combinator | https://www.ycombinator.com/launches/OAh-whispering-local-first-open-source-speech-to-text-at-your-fingertips

6. [Crunchbase, 2026] Epicenter - Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/epicenter-fab4

7. [Techno-Pulse, April 2026] Best AI Note-Taking Apps in 2026: Notion AI vs Mem vs Reflect vs Obsidian AI | https://www.techno-pulse.com/2026/04/best-ai-note-taking-apps-in-2026-notion.html

8. [GitHub, 2026] epicenter/apps/whispering at main · EpicenterHQ/epicenter | https://github.com/EpicenterHQ/epicenter/tree/main/apps/whispering

9. [Slator, 2026] Meet Whispering, an Open‑Source, Local‑First Transcription App - Slator | https://slator.com/whispering-open-source-local-first-transcription-app/

10. [GitHub, 2026] braden-w (Braden Wong) · GitHub | https://github.com/braden-w