ibl.ai Agentic AI Blog

Insights on building and deploying agentic AI systems. Our blog covers AI agent architectures, LLM infrastructure, MCP servers, enterprise deployment strategies, and real-world implementation guides. Whether you are a developer building AI agents, a CTO evaluating agentic platforms, or a technical leader driving AI adoption, you will find practical guidance here.

Topics We Cover

Featured Research and Reports

We analyze key research from leading institutions and labs including Google DeepMind, Anthropic, OpenAI, Meta AI, McKinsey, and the World Economic Forum. Our content includes detailed analysis of reports on AI agents, foundation models, and enterprise AI strategy.

For Technical Leaders

CTOs, engineering leads, and AI architects turn to our blog for guidance on agent orchestration, model evaluation, infrastructure planning, and building production-ready AI systems. We provide frameworks for responsible AI deployment that balance capability with safety and reliability.

Back to Blog

What Is a Poppy File, and What Should a Company Publish?

Miguel AmigotNovember 19, 2026
Premium

Sierra and Meta released Personal Agent Protocol 1.0 on November 13, 2026, and Sierra says every Shopify store now publishes a Poppy file. The file is a short list of pointers to a company's sign-in, APIs and agent, so what matters is who runs the systems it points to.

The Short Answer

A Poppy file is the JSON document a company publishes at /.well-known/poppy.json so a customer's personal AI agent knows how to sign in, what it may do and which APIs or company agent to call. The file is easy; the systems it points to are the work. On ibl.ai you own all the code and the data behind those endpoints, on any model.

The name comes from Personal Agent Protocol (Poppy), which Sierra and Meta are developing with more than 100 design partners. Version 1.0 shipped on November 13, 2026.

What did Sierra and Meta release on November 13, 2026?

On November 13, 2026, Bret Taylor and Clay Bavor announced Personal Agent Protocol 1.0 on Sierra's blog. The protocol, nicknamed Poppy, defines how a personal agent acting for one customer connects with a company.

Sierra says version 1.0 reflects about five weeks of design partner testing against real websites, APIs and agents. It adds 60 design partners, bringing the total to more than 100.

The draft came first. On October 9, 2026, Sierra published the draft with initial partners Meta, Genesys, Instinct, NiCE, Rocket, Shopify, Stripe and Walmart, plus 35 more design partners.

Does every Shopify store really publish a Poppy file now?

According to Sierra, yes. Its 1.0 announcement says "Shopify has enabled Poppy for every merchant on its platform" and that every Shopify store "now automatically publishes a Poppy file". The same post puts Shopify in more than 175 countries.

That is Sierra's account of a partner's rollout, not a Shopify filing. The way to confirm it for one store is to request that store's /.well-known/poppy.json.

When we checked one Shopify store, Allbirds, on November 18, 2026, its file declared protocol_version 0.4, not 1.0, and listed a ucp extension. Publishing a file and publishing the 1.0 version are not the same thing.

Shopify also co-authored Poppy's commerce extension. Sierra describes the split this way: Shopify's Universal Commerce Protocol (UCP) handles finding products, carts and checkout, and Poppy handles the step before, which is who the agent is and what it is allowed to do.

On the agent side, Sierra says Poppy is on for everyone who uses Muse, Meta's AI agent, which "automatically looks for Poppy first on every site it visits".

What does a Poppy file actually contain?

The Poppy 1.0 specification puts the file at https://{domain}/.well-known/poppy.json, served over HTTPS. It is a short list of declarations, and nearly every one is a pointer to a system the company has to run.

Poppy file section What it declares The decision behind it
auth Sign-in issuer, scopes, and hints on which scopes make irreversible changes or touch personal data Which identity system vouches for the customer, and what an agent may never do without asking
agent The company agent's protocols and HTTPS endpoints, with an optional access hint Whether a customer's agent talks to your agent, and whose agent that is
apis A list of APIs, each typed openapi or mcp, with a URL and description Which of your systems are exposed to agents at all
web A browser-session endpoint, or a flag asking personal agents not to browse Whether agents use the website like a person or go through APIs
tasks An endpoint for excluded tasks, skills and feedback What you refuse to let agents do, stated up front
extensions Supported extensions by name and major version, such as the ucp extension Shopify stores list Which industry rules you follow without changing the core protocol

Sierra's announcement says the protocol is built largely on existing standards, MCP, OpenAPI and UCP, with OAuth for sign-in, and that companies can adopt it in stages.

How is a Poppy file different from llms.txt or agents.md?

llms.txt and agents.md describe a site in prose so a model can read it. Shopify's own agents.md template documentation says a store serves agents.md at /agents.md, and that /llms.txt and /llms-full.txt mirror its content by default.

A Poppy file does a different job. It is not a description for reading; it is a contract for acting. It tells an agent how a customer signs in, which permissions exist, and which endpoints to call on that customer's behalf.

A company that wants to be both found and used by agents publishes both kinds. ibl.ai's own site serves an llms.txt for the first job; the second job depends on the systems behind the endpoints.

Why does it matter who runs the systems a Poppy file points to?

Because the file is mostly addresses. The sign-in issuer, the scopes, the MCP and OpenAPI endpoints and the company agent named in it are where a customer's agent is actually admitted or refused, and where every action is recorded.

If those endpoints belong to a vendor's managed service, the vendor's system answers the customer's agent. If they are yours, your rules and your logs do. The same question sits under Know Your Agent in banking: a company has to answer for agents from its own records.

This is where ibl.ai fits. Agentic OS connects a company's data to its agents over MCP, and you own all the code and the data, run it model-agnostic, and deploy it in your cloud, on-premise or air-gapped.

Pricing matters at this layer too. Personal-agent traffic is not tied to employee headcount, so a per-seat license is the wrong shape for it. ibl.ai is priced by usage or a flat license, with no per-seat pricing.

What should a company do about Personal Agent Protocol now?

Start with the decisions in the table, not the JSON. Sierra says companies can adopt Poppy in stages, and the 1.0 announcement links an implementation kit with a validator.

A practical order: pick the identity system that will issue sign-in for agents; mark which actions are irreversible; choose which systems to expose as MCP or OpenAPI; decide whether a company agent should answer at all.

Then check what you already publish. A Shopify merchant, by Sierra's account, already has a file; a company on its own stack has to write and serve one. Either way, the file can only promise what the systems behind it enforce.

Why does owning the AI stack matter?

ibl.ai is the agentic AI platform where you own all the code and the data. You self-host the entire stack inside your own perimeter, run it model-agnostic across any LLM and switch anytime, and pay by usage with no per-seat pricing β€” so you can deploy anywhere: your cloud, on-premise, GovCloud, or fully air-gapped.

  • You own all the code and the data

    Full source code under a perpetual license, running on your infrastructure. Not API access to someone else's platform β€” the stack itself is yours.

  • Model-agnostic

    Run any LLM β€” Claude, GPT, Gemini, Llama, Command, or your own fine-tune β€” and switch providers without rewriting the platform.

  • No per-seat pricing

    Usage-based billing against a budget cap you set. Cost tracks what your organization actually uses, not how many people you employ.

  • Deploy anywhere

    Your cloud, your VPC, on-premise, GovCloud, or a fully air-gapped network with no outbound connectivity.

1.6M+ users across 400+ organizations run the platform this way, including NVIDIA, MIT, and Syracuse University.

ibl.ai is family-owned and operated from New York, NY β€” a U.S.-headquartered, domestically-owned long-term partner, not a vendor that sells licenses and moves on.

See the ibl.ai AI Operating System in Action

Discover how leading universities and organizations are transforming education with the ibl.ai AI Operating System. Explore real-world implementations from Harvard, MIT, Stanford, and users from 400+ institutions worldwide.

View Case Studies
Work with our team

Pilots, deployment, and full ownership

Every engagement is scoped to the work: you integrate ibl.ai with your own data, deploy it on your own infrastructure, and the engineering hours scale with the scope β€” so the price tracks the scope, not your headcount. Most customers keep us on afterwards for platform updates, new agents and integrations, and ongoing services.

Start here

Pilot

from $15K

fixed scope Β· fixed timeline

A time-boxed proof of value on your real data β€” not a slide deck.

Best for: Teams that want to see ibl.ai working before committing.

  • Deployed on your infrastructure or our cloud
  • 1–2 production agents wired to a slice of your data
  • One integration (LMS / SIS / SSO / data source)
  • Weekly working sessions with our engineers
  • Pilot fee credits toward a full engagement
Scope a pilot
Most common

Integration & Deployment

$25K – $80K

one-time fee Β· scoped to the work

Full deployment integrated with your data and systems. Engineering hours scale with scope.

Best for: Organizations rolling ibl.ai out across a department, campus, or business unit.

  • Platform deployed in your VPC, on-prem, or air-gapped
  • Integrated with your data + identity (SSO / SAML)
  • Multiple custom agents built to your workflows
  • Engineering hours proportional to scope
  • You own the data Β· run any LLM you choose
Plan a deployment
Full ownership

Codebase Transfer + Custom AI Engineering

Custom quote

perpetual license Β· you own the stack

We transfer the full source code. You own and self-host the entire platform β€” outright.

Best for: Organizations and enterprises that benefit from perpetual ownership and sovereignty.

  • Complete source-code transfer + perpetual license
  • Dedicated AI engineering team on your roadmap
  • Custom agents, models, and integrations to spec
  • Air-gapped capable Β· zero vendor lock-in
  • Family-owned, New York–based long-term partner
Talk about ownership
You own the code and data Run any LLM β€” Claude, GPT, Gemini, Llama Family-owned & operated from New York, NY