Hypermode – The AI development platform
"Istari Digital Acquires Dgraph to Strengthen Data Foundation for AI and Engineering"
Value Proposition & Features
Hypermode is an AI development platform that provided hosted infrastructure for integrating AI models and data into applications, combining model and function hosting with observability, debugging, and automated deployment from GitHub.[16][121] It was positioned as infrastructure for developers shipping AI systems, not an end-user assistant, and is now documented as “shut down · historical record” with its main URL redirecting to a Dgraph acquisition notice.[17][99] Hypermode’s core value was enabling developers to connect application code to models and data sources while keeping credentials and execution isolated via a WebAssembly-based framework.[16][61][121]
Core product components:
- Hosted AI application infrastructure: Hypermode’s platform hosted AI models and functions, provided logs across separate executions, and supported debugging and automated deployments from GitHub, so developers could ship AI-backed applications with built-in observability.[16][121]
- Modus framework for agentic systems: Hypermode created Modus, an open-source serverless framework for AI applications and agentic systems written in Go or AssemblyScript, compiling functions into sandboxed API endpoints on WebAssembly and managing model/data access without exposing credentials.[16][61][91][121]
- Execution logging and behavior investigation: Customer-facing logs allowed teams to investigate behavior across distinct executions of their AI functions, connecting application events to AI-side activity.[16][121]
- Flexible hosting: Modus could run on developers’ own hardware or third-party hosting platforms, separately from Hypermode’s own hosted tooling.[16][121]
- Social campaign execution layer (later use of the name): Separately, Innovid’s Hypermode is described as a “high-precision execution layer” and “campaign-build interface for social” that supports Meta, Pinterest, Reddit, Snapchat, and TikTok and cuts campaign setup time by up to 80%; this appears to be a distinct product using the same name under Innovid/Mediaocean rather than the original AI developer platform company.[62][63][80][96]
Key features (historical Hypermode developer platform):
- Model and function hosting for AI apps — Hosted infrastructure for integrating AI models and functions into applications, with deployment from GitHub.[16][121]
- WebAssembly-based sandboxing (Modus) — Compiles functions into sandboxed API endpoints based on WebAssembly, separating application code from model and data credentials.[16][61][91][121]
- Observability and debugging — Customer-facing logs and execution traces help developers investigate behavior across runs of their AI systems.[16][121]
- Automated deployment from GitHub — Automated deployment pipeline connecting GitHub repositories to hosted functions and models.[16][121]
- Support for agentic systems — Modus is explicitly described as a framework for “agentic systems” as well as AI applications.[16][91][121]
- Self-hosted or external deployment — Modus can be run on developers’ own hardware or alternative hosting platforms, independent of Hypermode’s infrastructure.[16][121]
- Developer-focused (not end-user) — Positioned as infrastructure for developers building AI systems, contrasting with end-user assistants like Notis.[17][99]
- Status: shut down / historical record — Public comparison pages state Hypermode’s product URL is now dead and the platform is no longer active.[17][99]
Screenshots
No reliable source found for official product UI screenshots beyond generic Open Graph image infrastructure references; available material describes OG image behavior but does not provide confirmed Hypermode-specific UI captures.[46][48][52][55]
Product Roadmap / Announcements
As of September 11, 2026,
- 2026-09-06 – Status documented as shut down: A comparison of AI automation tools lists “Hypermode – AI development platform – Shut down · historical record”, noting that hypermode.com now only shows “Hypermode (fka Dgraph)” linking to Istari Digital’s acquisition announcement.[17][76][99]
- 2025-09 – Modus archived: Hypermode’s Modus repository is reported as “archived in September 2025 and read-only”, effectively freezing further framework development.[61][170]
- 2024 – Year-end company update: A 2024 retrospective notes Hypermode’s first customers entering production and highlights raising over $9 million that year, implying active roadmap and growth at that time.[61][91][136]
Recent Developments (past 90 days)
- 2026-09-06 – Third-party catalog marks Hypermode as shut down: Notis’s automation tools comparison page describes Hypermode as shut down and a historical record, with the production URL no longer serving the platform.[17][76][99]
- 2026-08-28 – Updated org profile summarizing archival status: An AI Engineer organization profile, refreshed August 28, 2026, documents Hypermode’s founding in 2023, its Modus framework, and notes the Modus repository was archived in September 2025, signaling the project’s end.[16][61][91][170]
- 2026-09-01 & 2026-08-12 – Innovid’s Hypermode execution layer news (distinct usage of name): PPC Land articles describe Hypermode within Innovid’s stack as a campaign-build and execution interface for social ads reaching general availability April 21, 2026, but this refers to Innovid’s product naming rather than the original hypermode.com AI developer platform.[62][63][95][96]
History and Origin Story
Hypermode was founded in 2023 as an AI application infrastructure company focused on a developer platform for integrating AI models and data into applications.[16][61][91][121] The company built Modus, an open-source, WebAssembly-based serverless framework for AI applications and agentic systems written in Go or AssemblyScript, allowing developers to compile functions into sandboxed API endpoints while keeping credentials hidden from application code.[16][61][91][121] A 2024 retrospective identified Kevin Van Gundy as CEO, reported raising over $9 million that year, and noted that Hypermode’s first customers were entering production and that its portfolio expanded through the acquisition of Dgraph Labs, although Dgraph was later sold to Istari Digital in October 2025.[61][91][136] The Modus repository was archived in September 2025, and subsequent profiles and comparison pages now describe Hypermode’s AI development platform as shut down, with hypermode.com redirecting to a Dgraph-related message.[61][170][17][99]
Fundraising History
| Round | Date | Amount | Lead investor |
| Seed / early funding (reported) | 2024 (year-end retrospective) | Over $9M raised in 2024 | Not publicly specified; described only as “raised over $9 million” in company update.[61][91][136] |
| Total | — | >$9M (reported 2024 funding) | Lead investor undisclosed in available sources.[61][91][136] |
Known investors (alphabetical):
- No specific investor names for Hypermode’s funding are disclosed in the available sources; only the aggregate funding amount is mentioned.[61][91][136]
Notable Team Members
Kevin Van Gundy (CEO): Hypermode’s 2024 year-end update identifies Kevin Van Gundy as CEO, noting his prior role as Chief Revenue Officer at DigitalOcean and placing him among key company sources referenced in the profile.[61][91][31][136]
Ryan Fox-Tyler: The Hypermode organization profile lists Ryan Fox-Tyler as an Amazon Web Services product leader associated with Hypermode, indicating a background in AWS product leadership that informed Hypermode’s infrastructure focus.[91]
William Lyon: The same profile cites William Lyon as a senior product manager, highlighting his role in shaping product direction for Hypermode’s AI application infrastructure and the Modus framework.[91]
Market Sizing
Category, Market Size, and Category Growth
Hypermode’s AI development platform and Modus framework fit within AI-powered software development platforms and agentic AI / agentic systems tooling, providing hosted infrastructure and a framework for agentic AI applications.[16][61][91][121][99] Market research estimates the global AI-powered software development platforms market at $1.8 billion in 2025, projected to reach $14.8 billion by 2034 at a 40.0% CAGR, while broader AI-native development platforms are valued at $10.0 billion in 2025 with a forecast of $204.09 billion by 2035 at 35.26% CAGR, indicating rapid growth in Hypermode’s category.[163][165] Agentic AI markets more broadly are sized around $7.29–11.56 billion in 2025–2026 with forecast compound annual growth rates generally in the 40–45% range across analyst firms, underscoring strong expansion for agentic AI tools and platforms.[151][153][154][159][161]
Pricing
| Tier / Model | Details |
| Custom enterprise pricing | A comparison of AI automation tools notes “Hypermode is infrastructure for developers shipping AI systems” and associates it with custom pricing, contrasting it with self-serve tools; however, specific price points are not disclosed.[76][99][17] |
No public pricing tables or per-seat/usage rates are available for Hypermode; sources indicate custom pricing with no published structure.[76][99][78]
Revenue Trajectory Estimates
Hypermode’s 2024 retrospective mentions over $9 million raised and the first customers entering production but does not disclose revenue or ARR figures.[61][91][136] No reliable source found for specific revenue, MRR, or ARR metrics for Hypermode’s AI development platform.
Competitive Landscape
Who it’s for, who it’s not for
Hypermode’s AI development platform was designed for software engineering teams and developers who needed hosted infrastructure, observability, and a framework to build, deploy, and debug AI-powered applications and agentic systems, particularly those comfortable with Go or AssemblyScript and WebAssembly-based sandboxing.[16][61][91][121] It suited organizations treating AI features as part of their application stack—requiring GitHub-driven deployment, logging across executions, and the ability to self-host the Modus runtime when desired.[16][61][121]
It was not aimed at non-technical end users or teams wanting a turnkey assistant, as comparison sources explicitly contrast Hypermode with products like Notis, describing Hypermode as infrastructure for developers rather than a ready-to-use assistant integrated into personal messaging and coding tools.[17][76][99] Similarly, buyers looking for low-code business user interfaces, pure marketing campaign tools, or pre-built vertical agents without developer ownership would be a poor fit relative to agent platforms and no-code automation products listed in broader market guides.[109][118][119]
Viable Alternatives
- LangGraph (framework and platform): Provides a graph-based framework and hosted platform for complex, stateful AI agent workflows, offering explicit state, branching, retries, and approvals suited to engineering teams; it occupies a similar “agent orchestration for developers” niche.[109][113][110][120]
- Amazon Bedrock AgentCore: Managed AWS-native infrastructure for production AI agents, combining runtime, identity, policy, memory, and observability, similar in spirit to Hypermode’s hosted infrastructure but deeply integrated into AWS.[110][111][112][118]
- Google Vertex AI Agent Builder / ADK: Google Cloud’s agent builder and Agent Development Kit (ADK) offer code-first plus managed services for multimodal agents with governance, making them close analogs for teams wanting hosted agent infrastructure.[110][111][112][118]
- Microsoft Agent Framework / Foundry Agent Service: Provides graph-based agent frameworks and managed Azure enterprise agent services with responsible AI guardrails, overlapping with Hypermode’s role as an AI development platform with infrastructure.[109][110][113][112]
- Relevance AI / LangGraph Platform (hosted): Commercial platforms focused on building and operating agents with observability, governance, and integrations, suitable for teams that might have considered Hypermode’s hosted tooling.[114][120][118]
Competitor Table
| Competitor | Description |
| LangGraph Platform | A code-first graph-based agent platform that lets engineering teams build complex, stateful AI workflows with checkpoints, retries, and human-in-the-loop controls, offering hosted and self-hosted deployment.[109][113][114][120] |
| Amazon Bedrock AgentCore | AWS-managed infrastructure for building and running production AI agents, covering runtime, identity, policies, memory, and observability for agents built on Bedrock models.[110][111][118] |
| Google Vertex AI Agent Builder | Google Cloud’s agent platform and ADK for constructing multimodal and text agents with production governance, integrating tightly with GCP services and data.[110][111][112][118] |
| Microsoft Agent Framework / Foundry Agent Service | Microsoft’s graph-based agent frameworks and managed Azure Agent Service for enterprise agents, providing built-in responsible AI guardrails and deep integration with the Microsoft ecosystem.[109][110][113][112] |
| Relevance AI | A low-code-plus-pro-code platform focused on building B2B SaaS agents and tools with observability, execution monitoring, and credit controls, suitable for teams seeking managed agent infrastructure.[114] |
| LangGraph (open-source framework) | An open-source agent orchestration framework providing explicit state and rollback for complex workflows, often deployed on third-party platforms similar to how Modus could be self-hosted.[109][113][118] |
| CrewAI | An open-source Python-based multi-agent framework oriented toward role-based automation, enabling teams of specialized agents, overlapping with Hypermode’s positioning around agentic systems.[109][113][115] |
| OpenAI Agents SDK | A tool-use loop and SDK for building simple agents with integrated tools, offering an alternative lightweight developer experience for agentic applications.[113][118][119] |
Sources
[1]: Hypermode: AI Application Infrastructure and Modus
[2]: AI Funding Tracker
[3]: AIR Raises $50M to Vet Skills and Add-Ons for AI Agents
[4]: Exclusive: Hypercubic raises $5.3M to modernize mainframe software
[5]: Hypercubic Raises $5.3 Million Seed To Turn Mainframe ... - Pulse 2.0
[6]: Hypercubic Raises $5.3 Million Seed To Turn Mainframe ... - LinkedIn
[7]: $100M Seed to Redefine How Humanity Computes - Callosum
[8]: Resources
[9]: Hypercubic Raises $5.3M Seed for Mainframe Modernization
[10]: HoneyHive raises $7.4M to bring evals and observability ...
[11]: Neuromorphic Labs Builds A 'trust Layer' For Business AI
[12]: Hyperbond Studio Raises $500K Seed Round
[13]: Higgsfield CEO Alex Mashrabov on new funding round, AI ...
[14]: Higgsfield Raises $400 Million Series B Financing at $5.4 Billion Valuation with Annualized Revenue Reaching $700 Million
[15]: Transfyr Launches With $25M in Seed Funding to Build ...
[16]: Compare AI automation tools side by side
[17]: Your agents can do more. You stay in control.
[18]:
[19]: testing-ink — Skill for Claude Code by LandonSchropp
[20]: Four new models for every job - Hyperagent
[21]: confluence-property — Skill for Claude Code by ulises-jeremias
[22]: planner — Cursor rule for Cursor by ulises-jeremias
[23]: tdd-guide — Agent for Claude Code by ulises-jeremias
[24]: agent-toolkit-agents — Plugin for Claude Code by ulises-jeremias
[25]: Skai and Innovid gain Reddit ad buying through completed API integrations
[26]: hydraia — Plugin for Claude Code by jdanigo
[27]:
[28]:
[29]: hyper-auto — Skill for Claude Code by zeikar
[30]: Jana Gajovská informace o Jana Gajovská (22.08.2026) - Rejstriky.finance.cz
[31]: LEON1DAS's Profile
[32]: Sim Shain's Post
[33]: Innovid gains Meta campaign data in NIVO through ads ... - PPC Land
[34]: Virtual Teams: Remote Team Management & Leadership - Coursera
[35]: August 13, 2026 - 10-Q: Quarterly report [Sections 13 or 15(d)]
[36]: Hypercharge Networks Q1 Loss Widens; Names Kyle Moncrief as CFO
[37]: Hyper Appoints Bernardo Perez as Vice President, Engineering
[38]: Our Leadership Team | TPC Leadership & Culture
[39]: 谢铭元 - 维基百科,自由的百科全书
[40]: Hyperion DeFi Inc (HYPD) (Q2 2026) Earnings Call ...
[41]: Leadership Team
[42]: Empathic Leadership Strategies for Effective Teams - LinkedIn
[43]: Since I'm pissed over the recent MGEX announcement. ...
[44]: Open Graph Images: What They Are & How They Work
[45]: Plane Docs | Plane
[46]: Blog — OG Images & Social Sharing
[47]: Dreamthreads | APIs.io Providers
[48]: opengraph-image and twitter-image in Next.js | Rune Hub
[49]: Open Graph Checker | Test Your Social Share Preview
[50]: satori - openai/plugins