Eve
A communications channel/tool mentioned for iMessage and WhatsApp support in personal shopper agents. It appears to provide mature messaging integrations for controlled agent workflows.
Key Highlights
- Eve was first shown as a minimal agent defined entirely in markdown and deployable with a single Vercel command.
- It was later described as a model-agnostic harness, placing it in the emerging category of full-fledged agent harnesses.
- Newsletter mentions tie Eve to durability infrastructure, observability, and runtime-agnostic deployment patterns.
- Its mature iMessage and WhatsApp channels make it especially notable for controlled personal shopper and messaging-based agent products.
At a glance
- Category
- Communications
Eve
Overview
Eve is a tool and agent harness discussed as a minimal, markdown-defined way to create and deploy agents, and later described as a model-agnostic harness with mature messaging channels for iMessage and WhatsApp. Across the mentions, it appears in two complementary roles: as an easy starting point for building agents and as infrastructure for controlled, durable, and observable agent workflows.For AI Product Managers, Eve matters because it sits at the intersection of agent authoring, deployment, and real-world communication channels. The mentions connect Eve to open-source and runtime-agnostic deployment options, ownership over runtime and model choice, and personal shopper agent experiences that require durable, encrypted messaging flows. That makes Eve relevant not just as a developer tool, but as part of product design for customer-facing AI agents.
Key Developments
- 2026-06-21: Guillermo Rauch demoed “eve” as a minimal agent defined entirely in markdown, using files like `instructions.md` and `skills/your-expertise.md`, and deployable with a single Vercel command.
- 2026-07-06: Harrison Chase cited EVE as part of the industry shift from agent frameworks toward full-fledged harnesses, alongside DeepAgents and Claude Agent SDK.
- 2026-08-29: Guillermo Rauch commented on Vercel’s announcement of Eve agent creation, positioning it as an easy way to start while contrasting it with approaches that provide deeper ownership of the Git repo, runtime, model choice, skills, tools, connectivity, and sandbox.
- 2026-09-12: Guillermo Rauch described Eve as a model-agnostic harness and pointed to ai-sdk.dev’s Harness Agent as a way to abstract across harnesses.
- 2026-10-21: Guillermo Rauch announced an open-source, runtime-agnostic solution extending the durability infrastructure behind Eve and Workflows to multiple harnesses, with local, Vercel, and DIY deployment support.
- 2026-11-21: Guillermo Rauch said mature Eve channels for iMessage and WhatsApp, together with improved models and better performance without sacrificing durability or encryption, now enable fully controlled personal shopper platforms on an open intelligence stack.
Relevance to AI PMs
1. Prototype and specify agents quickly: Eve’s markdown-based setup suggests a lightweight way to define agent behavior and skills, which can help PMs turn product requirements into testable agent specifications without heavy upfront engineering complexity.2. Design for channel-based customer experiences: The mentions of mature iMessage and WhatsApp support make Eve especially relevant for PMs building conversational commerce, concierge, or personal shopper experiences where the channel itself is part of the product.
3. Evaluate controllability and deployment tradeoffs: Eve is repeatedly associated with ownership of runtime, model choice, durability, observability, encryption, and deployment flexibility. PMs can use it as a reference point when deciding between simple hosted agent builders and more controlled open-stack architectures.
Related
- Vercel: Closely linked to Eve through deployment and agent creation workflows; one mention describes deploying Eve with a single Vercel command.
- Guillermo Rauch: The most frequent source in these mentions, framing Eve around markdown-based agent creation, runtime ownership, durability, and messaging channels.
- Markdown: Central to Eve’s early positioning as a minimal agent defined through markdown files like `instructions.md` and skill documents.
- Harrison Chase: Helped contextualize Eve as part of the market shift from frameworks to harnesses.
- DeepAgents: Mentioned alongside Eve as another example of a full-fledged harness.
- Claude Agent SDK: Also grouped with Eve in the harness category, useful for comparison when evaluating agent infrastructure choices.
- Harness Agent (ai-sdk.dev): Shared as a way to abstract across harnesses, reinforcing Eve’s role in a broader harness ecosystem.
- Workflows: Mentioned together with Eve in connection with shared durability infrastructure.
- agentcardhq: Appears in the personal shopper platform context that also highlighted Eve’s mature messaging channels.
Frequently asked questions
What is Eve?
A communications channel/tool mentioned for iMessage and WhatsApp support in personal shopper agents. It appears to provide mature messaging integrations for controlled agent workflows.
What should AI product managers know about Eve?
Eve was first shown as a minimal agent defined entirely in markdown and deployable with a single Vercel command. It was later described as a model-agnostic harness, placing it in the emerging category of full-fledged agent harnesses. Newsletter mentions tie Eve to durability infrastructure, observability, and runtime-agnostic deployment patterns.
Explore this topic
Newsletter Mentions (6)
“Guillermo Rauch said improved models, @agentcardhq, mature @eve iMessage and WhatsApp channels, and better performance without sacrificing durability or encryption now enable builders to ship fully controlled personal shopper platforms on an open intelligence stack.”
#18 𝕏 Guillermo Rauch said improved models, @agentcardhq, mature @eve iMessage and WhatsApp channels, and better performance without sacrificing durability or encryption now enable builders to ship fully controlled personal shopper platforms on an open intelligence stack. He recommended trying and customizing the personal shopper agent template.
“Guillermo Rauch announced an open-source, runtime-agnostic solution that extends the durability infrastructure powering Eve and Workflows to harnesses including Pi and OpenCode, with support for local, Vercel, or DIY deployments.”
#20 𝕏 Guillermo Rauch announced an open-source, runtime-agnostic solution that extends the durability infrastructure powering Eve and Workflows to harnesses including Pi and OpenCode, with support for local, Vercel, or DIY deployments. He added that Claude Code and Codex can be made durable and observable this way, though their lack of embedding support makes them less efficient and portable and requires an extra sandbox compared with Pi or Fx. Also covered by: @Guillermo Rauch
“Guillermo Rauch said Eve is a model-agnostic harness and shared ai-sdk.dev’s Harness Agent for abstracting harnesses.”
#11 𝕏 Guillermo Rauch said Eve is a model-agnostic harness and shared ai-sdk.dev’s Harness Agent for abstracting harnesses. #12 in Peter Yang expressed skepticism about “software factories,” arguing that—beyond verification and testing—AI cannot improve products or build features end-to-end without human involvement.
“Guillermo Rauch commented on Vercel’s announcement of eve agent creation, contrasting easy agent-building options with approaches that provide a Git repository and ownership of runtime, model choice, skills, tools, connectivity, and sandbox.”
#7 𝕏 Guillermo Rauch commented on Vercel’s announcement of eve agent creation, contrasting easy agent-building options with approaches that provide a Git repository and ownership of runtime, model choice, skills, tools, connectivity, and sandbox. He described Vercel’s linked page as the easiest way to start with eve.
“Harrison Chase observes the agent industry pivoting from frameworks like LangChain, AI SDK, and LlamaIndex to full-fledged harnesses such as DeepAgents, Claude Agent SDK, and EVE—with DeepAgents predating EVE by about ten months.”
#2 𝕏 Harrison Chase observes the agent industry pivoting from frameworks like LangChain, AI SDK, and LlamaIndex to full-fledged harnesses such as DeepAgents, Claude Agent SDK, and EVE—with DeepAgents predating EVE by about ten months.
“Guillermo Rauch demos “eve,” a minimal agent defined entirely in markdown (an instructions.md plus a skills/your-expertise.md) and deployable with one Vercel command.”
#1 𝕏 Guillermo Rauch demos “eve,” a minimal agent defined entirely in markdown (an instructions.md plus a skills/your-expertise.md) and deployable with one Vercel command. He argues markdown is set to become the next hot, most accessible programming language ever.
Related
Founder and CEO of Vercel, mentioned for discussing AI infrastructure, open intelligence stacks, and audited AI governance. He appears as a product and platform advocate for builders.
Founder/leader associated with LangChain, referenced for discussing LangSmith and tool gateway capabilities. He is a prominent AI infrastructure voice for agents and observability.
A developer platform company known for frontend and AI infrastructure. Here it is discussed in the context of AI infrastructure governance and auditing under ISO 42001:2023.
An agent framework that now supports dynamically loading tools when a skill is loaded. This pattern helps keep toolsets modular and can preserve prompt cache efficiency.
Anthropic's SDK for building Claude-powered agents and workflows. In the newsletter it is used in a formal verification example with Lean and TLA+.
Stay updated on Eve
Get curated AI PM insights delivered daily — covering this and 1,000+ other sources.
Subscribe Free