September 25, 2026 Vorim AI Team 4 min read
Vorim AI Has Signed the Cloud Security Alliance's AI Trustworthy Pledge We have signed CSA's AI Trustworthy Pledge. A pledge is a promise, not a certification, so here is how Vorim keeps each of its four principles with evidence you can check.
September 25, 2026 Vorim AI Team 6 min read
The Blueprint Alliance Names Four Open Standards. Vorim Now Ships All Four. The Blueprint Alliance whitepaper names four open standards its members validate against. This week Vorim shipped the last two, HTTP Message Signatures and Shared Signals with CAEP, and tested them against someone else's code.
September 24, 2026 Vorim AI Team 5 min read
From Authorised Agent Actions To Provable Outcomes When an agent moves money, two questions follow. Was it allowed to, and does what happened match what was approved? Vorim answers the first. We have built a proof of concept with Zeko Labs that answers the second, without putting a chain anywhere near verification.
September 20, 2026 Vorim AI Team 3 min read
Vorim AI Has Joined UKAI UKAI is the trade association for UK AI businesses, representing over 250 companies and carrying their views into government. We have joined as members, because the rules being written for agents right now will decide what British AI companies can build, and the people building should be in that room.
September 20, 2026 Kwame Nyantakyi 6 min read
Two New Integrations: Meta Muse and TypeSafe AI Meta's Muse agent and TypeSafe AI's Jev decision model both landed this month, and both put agents closer to actions that matter. Vorim now gates and signs custom tool calls on Muse Spark, and turns Jev's calibrated confidence into a policy floor with a signed record behind every decision. Both live in the TypeScript and Python SDKs from 3.21.0.
September 10, 2026 Vorim AI Team 4 min read
From Code to Motion: Vorim AI and SnappyCompliance Partner on Provable Physical AI AI agents are starting to move machinery, and the question changes from "was the agent allowed?" to "can we prove what happened, from the decision in software to the motion in the world?" Vorim signs the digital decision; SnappyCompliance covers the physical and hardware safety layer. Together: one continuous, auditor-ready record from code to real-world motion.
September 10, 2026 Vorim AI Team 5 min read
Corpus: Shared Memory That Learns Your Business, One Signed Entry at a Time Agents decide on context, and most teams feed it to them from a crawl of the file shares or a memory database anyone can write to. Vorim Corpus is the memory layer of the Multi-Agent Fabric: knowledge written by verified agents as they work, signed at the source, read under permission at the moment of decision. Here is what it does and where it sits in your stack.
September 6, 2026 Kwame Nyantakyi 5 min read
Vorim Now Guards Anthropic’s Commerce Agents Anthropic’s commerce-agents blueprint gives you a shopping agent that fills carts and hands off to checkout, and a merchant agent that reprices a live store. Its built-in gates are real but unsigned and in-process. Vorim now wraps both backends: every read, cart write, checkout handoff and live store change is permission-checked and signed, and the irreversible apply_change waits for a named human whose approval is on the record.
August 27, 2026 Kwame Nyantakyi 5 min read
Can You Prove What Your Agent Did to Someone Who Doesn’t Trust Your Logs? Most teams can see what their agents did. Far fewer can prove it to an auditor, a regulator or the other side of a dispute, because the record is kept by the people who might be at fault. Vorim ActionProof signs each agent action as it happens, so an outsider can verify it offline without taking your word for it.
August 20, 2026 Kwame Nyantakyi 4 min read
One Accountability Layer Across Every Model: Vorim and OpenRouter OpenRouter lets an agent route across 500+ models from 60+ providers through a single endpoint, swapping Claude for GPT for Llama on cost or fallback. The model underneath keeps changing. Vorim now sits over the top, so whichever model runs, every tool call the agent makes is permission-checked and signed into a record you can show an auditor.
August 18, 2026 Kwame Nyantakyi 5 min read
Answering a Prompt Before It Runs: Vorim and Claude Inference Hooks Anthropic quietly moved AI governance to a new place: the prompt itself, before the model runs. Claude Enterprise can now route every prompt to an external server for an allow-or-deny verdict, and a blocked prompt never reaches the model. Vorim is now that server — and unlike a plain filter, every verdict it returns is signed into a tamper-evident record you can show an auditor later.
August 12, 2026 Kwame Nyantakyi 4 min read
Watch the Session, Tighten the Agent: Two New Things A single agent is easy to reason about. A team of them handing work back and forth is where you lose the thread. We shipped two features for that: live multi-agent sessions, so you can watch a hand-off chain act in real time, and capability profiles, which look at what an agent actually did and tell you the smallest set of permissions it really needs.
August 12, 2026 Kwame Nyantakyi 5 min read
Guarding the Agent That Holds the Keys: Okta, Entra, and AWS The moment you let an agent administer your identity provider, it can create users, grant admin roles, and mint credentials. The token it holds usually inherits an admin's full power, and nothing proves which agent made which change. Vorim now sits in front of Okta, Microsoft Entra ID, and AWS IAM: it binds every privileged write to a verified agent, forces a human to approve the dangerous ones, and signs the record.
August 8, 2026 Kwame Nyantakyi 6 min read
Cloudflare Is Building the Agentic Internet. Vorim Is the Accountability Layer. Across a week of announcements, Cloudflare laid out its plan for an internet built for AI agents: a browser for agents, temporary accounts, an access model, Web Bot Auth, a leaner MCP. It is good, open, and mostly right. It also keeps stopping at the same line: proving which agent, acting for which human, did what, and was allowed to. That line is where Vorim lives, and this week we shipped two integrations to sit directly on top of it.
August 7, 2026 Kwame Nyantakyi 5 min read
Your Personal Agent Needs Trust Too Trust for AI agents is not just an enterprise problem. If you have ever asked a coding assistant to build you an agent, or run one that reads your email, books things, or spends your money, that agent is acting as you. Vorim gives the agents you build a scoped identity, a permission checked before they act, and a signed record you can actually see, and it plugs straight into the tools you already use.
August 6, 2026 Kwame Nyantakyi 5 min read
Proving a Real Human Is Behind the Agent Vorim can prove which agent took an action and whether it was allowed to. What it could never prove on its own is that a real person stands behind that agent. Our new TruAnon integration closes that gap, so before an agent does something that matters you can require that a genuine, anchored human is behind the request, and capture both identities in one signed record.
August 5, 2026 Vorim AI Team 5 min read
Your Company’s Real Knowledge Is Not in a Database The knowledge that actually makes a company good at its job lives in people’s heads and daily habits, not in a wiki. As agents start doing that work, they need somewhere to keep what they learn. The hard part was never storage. It is trusting what is in the pool. Here is how we think about turning working knowledge into a governed asset, and the piece of it Vorim already ships.
August 3, 2026 Vorim AI Team 6 min read
When Agents Start Paying: Identity and Signed Audit for Every Payment Rail AI agents have stopped just recommending and started acting. They check out, they move money, they send the payment. That surfaces a question almost no one built for: which agent paid, was it allowed to, and can you prove what it did afterward? Vorim now covers eleven agentic-payment rails with a single authorization step and a signed record of the whole payment lifecycle.
July 30, 2026 Vorim AI Team 7 min read
The Agent Trust Scorecard: Are Your Agents Safe to Run in Production? Most teams can tell you which model their agent runs on. Far fewer can tell you what happens when that agent does something it should not. Our new free scorecard grades your deployment on the ten things that actually keep agents trustworthy, safe, and secure, drawn from OWASP, NIST, the CSA, and the EU AI Act.
July 26, 2026 Vorim AI Team 6 min read
The Multi-Agent Fabric: A Trust Layer for Agents That Work Together Agents rarely work alone anymore. One hands a task to another, they share context, and they act on your behalf as a team. The frameworks that orchestrate them assume they already trust each other. Vorim now gives a team of agents a shared trust layer: verified hand-off, governed shared memory, and a provable record of the whole workflow, so collaboration never becomes a black box.
July 22, 2026 Vorim AI Team 4 min read
AI Cost Management: Token Budgets That Stop a Runaway Agent Before the Invoice An autonomous agent that loops, retries, or gets stuck can burn through tokens faster than anyone is watching. Vorim now lets you set a rolling token budget per agent or per workflow, enforced at the decision point, so the one workflow that runs away is blocked before the spend, not discovered on next month’s bill.
July 21, 2026 Vorim AI Team 3 min read
Vorim AI Joins the Databricks Connected Partner Program Vorim AI has been accepted into the Databricks Connected Partner Program. Enterprise AI agents increasingly act on the data platform, querying warehouses and touching production data. That is exactly where the question "which agent did this, and was it allowed to?" gets sharpest.
July 15, 2026 Kwame Nyantakyi 6 min read
The Agentic Web Is Arriving. Every Agent Will Need an Identity, Permissions, and an Audit Trail. Analysts expect over a billion AI agents deployed worldwide by 2029, and the average Fortune 500 to run more than 150,000 agents by 2028. As autonomous agents move from demos into production, someone has to answer which agent did what, and whether it was allowed to. Here is why that becomes the defining infrastructure problem of the agentic web, and how Vorim solves it.
July 6, 2026 Kwame Nyantakyi 3 min read
Vorim AI Joins the NVIDIA Inception Program Vorim AI has been accepted into NVIDIA Inception, the program that nurtures startups building at the cutting edge of AI. It is a strong signal for where we are headed: the identity and trust layer that lets enterprises deploy AI agents in regulated, high-stakes environments with confidence.
June 26, 2026 Kwame Nyantakyi 6 min read
Delegation Receipts: Proving an AI Agent Only Ever Lost Authority, Offline Every agent identity vendor answers "is this agent allowed?" by checking a token against their server. Take the server away and the proof evaporates. @vorim/verify 0.5.0 ships delegation receipts: a human-readable, offline-verifiable record showing exactly how authority narrowed at each hop of a delegation chain. No vendor. No blockchain.
June 25, 2026 Kwame Nyantakyi 2 min read
Vorim AI Joins Claude for Startups and OpenAI for Startups Vorim AI has been accepted into both Claude for Startups and OpenAI for Startups. A short note on what it means for the work, and a thank-you to the people building alongside us.
June 23, 2026 Kwame Nyantakyi 7 min read
Let Your Agent Set Up Vorim: Agent-Led Onboarding With the Device Authorization Grant An AI coding agent can now onboard you to Vorim itself, getting a scoped, server-clamped API key after you approve once in the browser. No key pasting. Here is how the device flow works, why we kept a human in the loop, and the one-call setup.
June 17, 2026 Kwame Nyantakyi 6 min read
The Identity and Audit Layer for Google's Universal Commerce Protocol (UCP) Google's UCP lets AI agents discover, check out, and pay across millions of merchants. It standardises the transaction. It does not answer which agent is spending, whether it is allowed to, or how a counterparty proves what it bought. We built that layer.
June 4, 2026 Kwame Nyantakyi 7 min read
What "Offline-Verifiable" Actually Means, and How to Demo It in Five Minutes Every agent identity vendor we tested says its audit trail is "tamper-evident." None of them ships an audit bundle a third party can verify without the vendor in the trust path. Here is what changes when one does.
May 6, 2026 Kwame Nyantakyi 8 min read
Why Agent Identity Should Not Ship with the Model When Microsoft launched Agent 365, the question on every founder's LinkedIn was "is this game over?" It's the wrong question. The right one: where does the trust layer for AI agents belong in the stack?
April 18, 2026 Kwame Nyantakyi 5 min read
Giving OpenClaw Agents Identity, Permissions, and Audit Trails OpenClaw agents can browse the web, execute shell commands, and send emails. Now they can prove who they are, stay within their permissions, and log every action with a tamper-proof audit trail.
April 16, 2026 Kwame Nyantakyi 6 min read
Securing Agent Commerce: Identity Verification for Stripe's Agentic Commerce Protocol AI agents can now buy things via Stripe ACP. But who verifies the agent before it spends money? We built the identity layer for agentic commerce.
April 15, 2026 Kwame Nyantakyi 6 min read
Type-Safe Agent Identity for Pydantic AI: Permissions and Audit Trails the Pydantic Way Pydantic AI gives your agents type safety. Vorim gives them identity, permissions, and audit trails. Here's how to combine them with dependency injection.
April 15, 2026 Kwame Nyantakyi 7 min read
Introducing @vorim/a2a: The Identity and Trust Layer for Google's A2A Protocol AI agents can now discover and talk to each other via Google's A2A Protocol. But how does Agent A know Agent B is trustworthy? We built @vorim/a2a to solve that.
April 11, 2026 Vorim AI Team 10 min read
Credential Delegation for AI Agents: How to Give Agents OAuth Access Without Sharing Secrets Your AI agents need to access Google Drive, GitHub, and Slack on behalf of users. But sharing refresh tokens directly is a security liability. Here's how credential delegation solves this — and how to implement it with Vorim AI.
April 11, 2026 Vorim AI Team 8 min read
Ephemeral Agents: Temporary Identity for Short-Lived AI Agents Using Portable Decentralized Identifiers Not every AI agent needs to live forever. Ephemeral agents use portable decentralized identifiers to get full cryptographic identity, permissions, and audit trails, then auto-expire when their job is done.
April 11, 2026 Vorim AI Team 9 min read
The Compliance Layer That Unlocks OpenAI Agents in Regulated Enterprise OpenAI agents are powerful. But enterprise security teams won't approve them without identity verification, permission scoping, and compliance-grade audit trails. Vorim AI provides the accountability layer that makes deployment possible.
April 11, 2026 Vorim AI Team 9 min read
The Trust Infrastructure That Makes Multi-Agent Claude Safe for Enterprise Anthropic built Claude to be helpful, harmless, and honest. But enterprise deployment of multi-agent Claude systems requires more than safety alignment — it requires verifiable identity, scoped permissions, and signed audit trails.
April 11, 2026 Vorim AI Team 8 min read
The Identity Layer for the Open-Source Agentic Ecosystem LangChain, CrewAI, LlamaIndex, AutoGen — the open-source agent ecosystem is thriving. But none of these frameworks ship with identity infrastructure. Vorim AI is the missing layer that makes open-source agents enterprise-safe.
April 11, 2026 Vorim AI Team 10 min read
An Open Identity Protocol for the Agentic Web The web needed SSL. APIs needed OAuth. Cloud needed IAM. The agentic web needs an open identity protocol. VAIP is that protocol — and it's designed to be steward-agnostic, standards-based, and infrastructure-grade.
April 11, 2026 Vorim AI Team 9 min read
AI Agent Identity and Accountability at Scale: Infrastructure for the Next Generation of AI Systems As AI agents scale from hundreds to millions, the infrastructure for identity, permissions, and audit trails must scale with them. Here's what accountability looks like at production scale — and why it matters for every AI company.
March 31, 2026 Vorim AI Team 12 min read
The Practical Guide to Securing AI Agents: Identity, Trust & Accountability AI agents are shipping to production without verifiable identity, scoped permissions, or tamper-proof audit trails. This guide covers why agent trust infrastructure matters, what regulations are driving it, and how to implement it in your stack today.
March 30, 2026 Vorim AI Team 10 min read
US AI Laws & Executive Order 14110: How to Build Compliant AI Agents in 2026 The US is regulating AI at both federal and state level. Executive Order 14110 sets the tone, and states like Colorado, Illinois, Texas, and California are writing the rules. Here's what AI agent developers need to know — and how to stay compliant.
March 25, 2026 Vorim AI Team 8 min read
Securing LangChain and OpenAI Agents in Production: A Practical Guide How to add identity verification, permission checks, and audit logging to LangChain and OpenAI agents using the Vorim SDK. Production-ready patterns with code examples.
March 22, 2026 Vorim AI Team 7 min read
EU AI Act Compliance for AI Agents: What You Need to Know in 2026 The EU AI Act is now enforceable. Here's what it means for teams deploying autonomous AI agents — and how identity, audit trails, and trust scoring keep you compliant.
March 12, 2026 Vorim AI Team 8 min read
Why AI Agents Need an Identity Layer — And Why It Can't Wait As autonomous agents move from demos to production, the lack of standardized identity infrastructure is becoming the single biggest blocker to enterprise adoption. Here's why identity is the missing layer.
March 10, 2026 Vorim AI Team 10 min read
Building a Trust Protocol for AI Agents: From Zero Trust to Earned Trust Zero Trust architecture assumes nothing can be trusted. But for AI agents to collaborate effectively, they need a path from zero trust to earned trust. Here's how we designed Vorim AI's trust scoring system.
March 6, 2026 Vorim AI Team 7 min read
Permission Design Patterns for Autonomous AI Agents How do you design a permission system for software that makes its own decisions? We share 5 battle-tested patterns from production agent deployments.
March 1, 2026 Vorim AI Team 6 min read
Cryptographic Audit Trails: How to Make AI Agent Actions Compliance-Ready Regulators are coming for AI. Signed audit bundles, cryptographic manifests, and tamper-proof event chains are how you prepare. A practical guide.
February 24, 2026 Vorim AI Team 12 min read
Towards an Agent Identity Standard: Why AI Needs Its Own RFC HTTP has RFCs. TLS has RFCs. OAuth has RFCs. AI agent identity has... nothing. It's time to change that. Here's our proposal for what an agent identity standard should look like.