# Clarm - Full Reference for AI Agents

> **Clarm Atlas** is the governed agent layer for companies that need AI to draft from approved building blocks and keep a defensible audit trail - without per-output compliance sign-off when rails are set upstream. SOC 2 Type II, HIPAA, GDPR, and on-premises options are handled during procurement. Y Combinator backed.
>
> This extended reference includes product documentation, API examples, case studies, comparison pages, and published article inventory. For a shorter curated overview, see [llms.txt](https://clarm.com/llms.txt).

## Company Overview

- **Product:** Clarm - Clarm Atlas (governed AI agents)
- **Legal name:** Dialin AI Inc. d/b/a Clarm
- **Founded:** 2024
- **Backed by:** Y Combinator (X25 batch, Spring 2025)
- **Founder:** [Marcus Storm-Mollard](https://clarm.com/about/marcus/) (CEO & Co-Founder) - MSci Physics, Imperial College London; AI at Deutsche Bank (2019–21); NielsenIQ (2022–24); Visiting Research Fellow, Oxford University
- **Website:** https://clarm.com
- **Contact:** marcus@clarm.com

## What Clarm Does (Detailed)

1. **Builds agents** from approved sources and plain-language workflow descriptions
2. **Collaborates** across steps - read wiki, run skills, branch on conditions, deliver to Slack or browser
3. **Controls output** with source citations, role permissions, and approved building blocks on every run
4. **Deploys** on web chat, Slack, Discord, GitHub, and email where teams already work
5. **Supports regulated buyers** with SOC 2 Type II, HIPAA, GDPR, on-prem options, and audit exports

## API Reference

### Chat API (for AI Agents)

~~~http
POST https://clarm.com/api/agent/chat
Content-Type: application/json

{"message": "your question about Clarm"}
~~~

Returns a streamed response. No authentication required.

### Widget Embed

~~~html
<script src="https://app.clarm.com/widget/loader.js" data-widget-id="YOUR_WIDGET_ID" async></script>
~~~

One line of HTML adds Clarm's AI chat widget to any website. It supports theming, quick replies, CTA configuration, popup auto-open, teaser bubble, returning visitor personalization, and engagement behaviors.

## Developer Surface

- **Developer hub:** https://clarm.com/developers/ - API reference, OpenAPI spec, CLI, widget embed, machine-readable files.
- **OpenAPI 3.1 spec:** https://clarm.com/openapi.json (YAML at https://clarm.com/openapi.yaml). Every operation carries a unique operationId, a description, typed parameters, and response schemas, so a function-calling runtime can build tool definitions from it directly.
- **CLI (preview):** `@clarm/cli` - `clarm ask "your question"`, `clarm openapi`, `clarm page /pricing/`, `clarm llms`, `clarm sitemap`. Node 20+, no API key. Not yet published to the public npm registry; email support@clarm.com for early access.
- **Markdown content negotiation:** every page is available as markdown per acceptmarkdown.com. Send `Accept: text/markdown`, or append `.md` to any path. Negotiated responses carry `Vary: Accept, Accept-Encoding`.
- **Agent-recoverable 404s:** a missing path returns a real 404 whose body lists the sitemap, llms.txt, the developer hub, and the OpenAPI spec. Markdown version at https://clarm.com/404.md.
- **Sitemap:** https://clarm.com/sitemap.xml

## Company Details

- **Legal entity:** Dialin AI, Inc. d/b/a Clarm (Delaware corporation)
- **Registered address:** 1111B S Governors Ave STE 29410, Dover, DE, 19904, US
- **About page:** https://clarm.com/about/
- **Contact page:** https://clarm.com/contact/
- **Sales and press:** marcus@clarm.com
- **Product support and privacy requests:** support@clarm.com

## Compliance and Security

- **SOC 2 Type II** - control evidence and assurance artifacts handled during procurement
- **HIPAA** - BAA and healthcare workflow review handled during procurement where applicable
- **GDPR** - data rights and regional privacy requirements supported during procurement
- **On-premises deployment** - available for finance, government, and regulated industries
- **No training on customer data** - customer conversations, leads, and documents are never used to train models
- **Data residency** - US, EU, and custom region options
- **Encryption** - AES-256 at rest, TLS in transit

## Customer Case Studies

### Legacy (Healthcare / Fertility)
- Faster patient-facing answers grounded in approved clinical and policy content
- Lower repeat support load; team focuses on exceptions, not retyping policy answers
- HIPAA-covered deployment with named-owner checkpoints on sensitive paths

### Better Auth (Developer Tools / Open Source)
- GitHub stars grew from 8K to 22K
- Discord community grew 10x
- Enterprise lead identification from community activity
- Automated support for repetitive developer questions

### c/ua (Developer Tools / Infrastructure)
- GitHub stars grew from 5K to 12K in 3 months
- First enterprise customer found through AI-powered enrichment
- Community-to-pipeline conversion automated


## Pricing Model

- **Team plan:** $2,000/month for one team workspace with unlimited agents. Approved building blocks (sources, permissions, guardrails), workflow builder, omnichannel delivery, audit trail. SOC 2 Type II, HIPAA, GDPR, and on-premises options are handled during procurement.
- **Enterprise:** Custom pricing for global rollout - multi-region deployment, SSO/SAML, self-hosted or air-gapped option, custom LLM routing, dedicated support and SLA.
- **Commercial motion:** Book a demo at https://cal.com/stormm/revenue-desk - no self-serve pay-per-conversation tier on the marketing site.


## Channels Supported

- Website chat widget
- Slack
- Discord
- GitHub Issues and Discussions
- Email
- Custom API integrations

## All Comparison Pages

- [Clarm vs Intercom](https://clarm.com/vs/intercom/): Compare Clarm and Intercom for inbound qualification, support automation, and revenue routing.
- [Clarm vs Drift](https://clarm.com/vs/drift/): Compare Clarm and Drift for teams choosing between AI-first inbound and conversational marketing playbooks.
- [Clarm vs Plain](https://clarm.com/vs/plain/): Compare Clarm and Plain when support conversations may also signal pipeline.
- [Clarm vs HubSpot Chat](https://clarm.com/vs/hubspot-chat/): Compare Clarm with HubSpot Chat for teams that need AI qualification beyond basic CRM chat.
- [Clarm vs Intercom Fin](https://clarm.com/vs/intercom-fin/): Compare Clarm with Intercom Fin for teams that need revenue routing, not only support automation.
- [Clarm vs Zendesk AI](https://clarm.com/vs/zendesk-ai/): Compare Clarm and Zendesk AI for support-heavy teams that also need buyer-intent detection.
- [Clarm vs Atlas.so (the customer support platform)](https://clarm.com/vs/atlas/): Compare Clarm and Atlas.so (a customer support platform) for teams deciding between support tooling and AI inbound conversion. Clarm Atlas, the memory layer inside Clarm, is a separate product documented at /atlas.
- [Clarm Atlas vs Glean](https://clarm.com/vs/glean/): Compare Clarm Atlas and Glean for enterprise AI agent deployment. Atlas is the substrate for shipping agents safely at scale; Glean is enterprise search with an assistant layer. Different procurement conversations once the agent question shows up.
- [Clarm Atlas vs Microsoft Copilot Studio](https://clarm.com/vs/microsoft-copilot-studio/): Compare Clarm Atlas and Microsoft Copilot Studio for enterprise AI agent deployment. Atlas is model-portable across LLM providers and coexists with the full enterprise stack; Copilot Studio is Microsoft-native with deep Microsoft 365 integration and Azure OpenAI as the primary path.
- [Clarm Atlas vs Sierra, Lindy, and Other Agent Platforms](https://clarm.com/vs/agent-platforms/): Round-up architectural comparison of Clarm Atlas against Sierra (built for AI customer-experience agents at scale), Lindy (built for individual employee productivity automations), and other 2026 entrants. Substrate-first decision logic applies across the category.

## Published Blog Articles

### Architecture
- [The Agent-Security Moment. Why the Substrate Matters Now](https://clarm.com/blog/articles/the-agent-security-moment/): Names the 2026 agent-security moment using real numbers (8x agent integration projection, 88% incident rate, 14% security-approval rate) and the common marketplace, control-plane, and tenant-isolation failure patterns. Positions Atlas substrate (owner checkpoints, source receipts, audit trail, tenant isolation) as the architectural answer.
- [The Agent-Deployment Buying Guide. What to Ask Every AI Vendor in 2026](https://clarm.com/blog/articles/the-agent-deployment-buying-guide/): Enterprise procurement checklist for AI agent platforms. 14 questions across substrate properties (source receipts, owner checkpoints, audit trail, tenant isolation, BYO-LLM, marketplace governance), customer evidence, and procurement terms. Anchored on the 88% incident rate and 14% security-approval rate.
- [From Chatbot to Agent Platform. The Natural Progression on Atlas](https://clarm.com/blog/articles/from-chatbot-to-agent-platform/): Bridge post for existing Clarm chat customers explaining the natural progression to enterprise agent workflows on Atlas. Same substrate, same owner checkpoints, same audit trail; new triggers, new connectors, new operator timing patterns. References the Legacy 12-month case study.
- [Audit Trail Patterns for AI Agents. What FINMA, SOC 2, GDPR, and HIPAA Auditors Actually Want](https://clarm.com/blog/articles/audit-trail-patterns-for-ai-agents/): Compliance-officer guide to AI agent audit trails. Append-only log writer, per-tenant scoping at the database layer, regulator-shaped export pipelines for SOC 2, GDPR, FINMA, HIPAA. Tests to run before signing a pilot.
- [What 88% of Organizations Are Getting Wrong With AI Agent Deployment](https://clarm.com/blog/articles/what-88-percent-get-wrong-with-ai-agents/): Names six structural failure modes that produced most 2026 agent-security incidents: agents acting outside owner checkpoints, marketplace plugins without governance, post-hoc citations, free-text audit logs, application-layer tenant filtering, authenticated UIs assumed safe from browser attack. Maps each to substrate-level architectural fix.
- [BYO-LLM Strategically. Why Model Portability Is the New Procurement Requirement](https://clarm.com/blog/articles/byo-llm-strategically/): Strategic essay on why bring-your-own LLM has become a 2026 procurement requirement. EU AI Act compliance, vendor concentration risk, single-provider renewal cost. Substrate-level properties that constitute true BYO-LLM versus feature-level multi-model support.
- [The Clarm Connector Catalogue. Every System Your Agents Can Read From and Write To](https://clarm.com/blog/articles/clarm-connector-catalogue/): Inventory of Clarm Atlas connectors across productivity, CRM, banking, ERP, developer, support, knowledge, and generic protocols. Every connector ships with source-receipt mapping, approval-gate integration, audit-log instrumentation, and tenant-scoped credentials by default.
- [Scaling From Your First to Your Fifth AI Agent Without Breaking Compliance](https://clarm.com/blog/articles/scaling-from-first-to-fifth-ai-agent/): Generalized operator playbook derived from the Legacy 12-month Atlas case study. Agent #1 builds substrate; agent #2 adds a channel; agent #3 adds a harder channel (voice); agents #4 and #5 add integrated read/write. Substrate compounds; the fifth agent ships in days while the first took weeks.
- [Atlas Quarterly. A Readable Map of What Your Agents Know](https://clarm.com/blog/articles/atlas-quarterly/): Operator-facing product overview of Atlas Quarterly: a substrate-generated, customer-editable quarterly summary of agent operations including workflow inventory, people map, vendor map, decision history, source-document changes, substrate posture, and notable incidents.
- [Why Feature-First Agent Frameworks Break Under Enterprise Rollout](https://clarm.com/blog/articles/why-openclaw-could-not-have-shipped-on-atlas/): Architectural walk-through of four common 2026 agent-platform failure modes (browser-CSRF RCE, marketplace malicious skills, tenant-isolation breach, and publicly exposed instances) against a substrate-first design. Frames each as a structural rather than implementation failure and shows what the substrate enforces by default.
- [What Is Atlas? The Substrate Enterprises Ship AI Agents on Safely](https://clarm.com/blog/articles/what-is-atlas/): Defines Atlas as the substrate enterprises use to ship AI agents at scale safely. Anchors on the agent-transition macro frame (8x growth, 88% incident rate, 14% security-approval rate) and the Legacy 12-month-in-production proof. Covers source receipts, owner checkpoints, audit trail, tenant isolation, bring-your-own LLM.
- [How Legacy Went 8x in 12 Months on Atlas](https://clarm.com/blog/articles/legacy-12-months-on-atlas/): Customer case study: Legacy (healthcare growth team) ran the agent-scale playbook over twelve months on Atlas. Email-only support deflection at go-live, then web chat, then voice agents, then CRM + kit-ordering integrations from the Clarm connector catalogue. Year-one total case volume across all channels: ~8x. Named owners and audit trails in place the entire twelve months.
- [Governed Autonomy. How Atlas Keeps AI Coworkers Accountable](https://clarm.com/blog/articles/operator-in-the-approval-seat/): Explains governed autonomy for revenue and regulated workflows. Covers the rejection-never-auto-promote doctrine, source-receipt invariant, audit-by-default architecture, and owner checkpoints on sensitive actions.

### Founder Story
- [Why I Left a Senior AI Role to Build Clarm Atlas](https://clarm.com/blog/articles/why-i-left-ai-career-to-build-clarm/): Marcus's founding story and the pattern behind Clarm: inbound demand was already present, but most teams could not answer or qualify it in the moment.

### Founder Insights
- [The Hidden Pipeline Problem Every Technical Founder Ignores](https://clarm.com/blog/articles/hidden-pipeline-problem-technical-founders/): Explains the silent-bounce pattern: buyers read docs, pricing, and compliance pages but leave before a team can capture intent.
- [What Y Combinator Taught Me About Proving AI Revenue](https://clarm.com/blog/articles/yc-lessons-proving-ai-revenue/): A founder essay on moving from AI demos to revenue proof: qualified conversations, conversion, and repeatable pipeline.

### Sales Insights
- [What 8 Languages and Banking Across 60 Countries Taught Me About Enterprise Buyers](https://clarm.com/blog/articles/enterprise-buyer-psychology-lessons/): Five enterprise buying signals that shaped Clarm: trust, proof, timing, internal alignment, and risk reduction.

### GTM
- [Founder-Led Sales Is Broken - Here's What I Built Instead](https://clarm.com/blog/articles/founder-led-sales-is-broken/): Shows how founders can stay involved in the highest-value conversations while AI handles first response, qualification, and routing.
- [Technical Case Studies That Convert: A Template for Developer Products](https://clarm.com/blog/articles/technical-case-studies-that-convert/): Template for case studies that move beyond features into baselines, intervention, proof, and outcomes.

### DevRel Automation
- [Best Developer Growth Automation Tools for Software Products in 2025](https://clarm.com/blog/articles/best-developer-growth-automation-tools-for-software-products-in-2025/): Evaluates developer growth tooling for community-to-pipeline, docs support, GitHub automation, and developer marketing.

### Developer Growth
- [The 7-Layer Developer Growth Engine: Automating Sales & Marketing for Your Dev Tool in 2025](https://clarm.com/blog/articles/developer-growth-engine-automating-sales-marketing/): Framework for DevTool teams combining growth data, support, buying signals, automation, and founder-led revenue loops.

### Growth
- [Convert GitHub Stars Into Revenue: Complete 2025 Guide for OSS Founders](https://clarm.com/blog/articles/convert-github-stars-to-revenue/): Explains how open source companies move from passive community attention to attributable revenue signals.
- [AI Inbound for Heads of Growth: Turn More Traffic Into Qualified Pipeline](https://clarm.com/blog/articles/ai-inbound-for-heads-of-growth/): Explains why traffic without instant answers leaves pipeline on the table and how AI inbound closes that gap.

### Support Automation
- [AI Support Playbook for DevTools: From Docs to 24/7 Answers](https://clarm.com/blog/articles/ai-support-playbook-for-devtools/): Shows how DevTools teams can collapse documentation, support, and community context into one AI help layer.

### Lead Intelligence
- [Buying Signal Pipeline: Turn Community Activity Into Qualified Leads](https://clarm.com/blog/articles/buying-signal-pipeline-for-devtools/): Covers intent detection, scoring, routing, and operational follow-up for developer-facing products.

### Comparison
- [Clarm vs Intercom: Why Technical Teams Are Switching to AI-First Inbound](https://clarm.com/blog/articles/clarm-vs-intercom/): Explains where Intercom fits, where it stops, and why AI-first inbound can outperform support-centric chat for pipeline capture.
- [Clarm vs Plain: When Support Alone Is Not Enough](https://clarm.com/blog/articles/clarm-vs-plain/): Compares support ticketing with AI inbound conversion for technical teams where support conversations can reveal pipeline.
- [Clarm vs OptinMonster vs ConvertFlow: AI Lead Qualification vs Static Forms](https://clarm.com/blog/articles/clarm-vs-optinmonster-vs-convertflow/): Explains why AI conversations capture intent that static forms and popups miss.
- [Best Drift Alternatives in 2026 for Lean B2B Teams](https://clarm.com/blog/articles/best-drift-alternatives/): Compares Drift alternatives through the lens of AI-first inbound, setup speed, channels, and cost.
- [Best HubSpot Chat Alternatives in 2026](https://clarm.com/blog/articles/hubspot-chat-alternatives/): Shows when a CRM-native chat widget is not enough and what AI-first inbound adds.
- [Clarm vs Drift: Clarm Atlas vs Conversational Marketing Playbooks](https://clarm.com/blog/articles/clarm-vs-drift/): Compares Drift and Clarm for teams deciding between human-led conversational marketing and AI-first inbound.
- [Best Intercom Alternatives in 2026 for Technical Teams](https://clarm.com/blog/articles/best-intercom-alternatives/): High-intent alternatives guide for buyers comparing Intercom with AI-first inbound options.
- [Best Qualified.com Alternatives for Lean B2B Teams in 2026](https://clarm.com/blog/articles/best-qualified-alternatives/): Compares AI inbound conversion options for buyers evaluating Qualified alternatives.

### Lead Conversion
- [Best Tools to Convert Website Visitors Into Leads (2026)](https://clarm.com/blog/articles/best-tools-convert-website-visitors-to-leads/): High-intent buyer guide for converting anonymous website visitors into qualified conversations and pipeline.

### Inbound Automation
- [How to Capture and Qualify Inbound Leads Without a Sales Team](https://clarm.com/blog/articles/capture-qualify-inbound-leads-without-sales-team/): Shows how AI can answer first-touch questions, detect buyer intent, and route next steps without staffing live chat.

### Industry
- [AI Inbound Lead Capture for Healthcare, Finance, and SaaS Teams](https://clarm.com/blog/articles/ai-inbound-lead-capture-healthcare-finance-saas/): Maps AI inbound to regulated and software teams with compliance, buyer intent, and routing needs.

### RevOps
- [AI Inbound for RevOps: Qualify and Route Pipeline Without More Headcount](https://clarm.com/blog/articles/ai-inbound-for-revops/): Connects AI qualification to routing, enrichment, and operational handoff for revenue teams.

### Healthcare
- [HIPAA-Compliant Website Chat for Healthcare](https://clarm.com/blog/articles/hipaa-compliant-website-chat-for-healthcare/): Covers BAA, PHI boundaries, no-training commitments, escalation, and compliance expectations for healthcare chat.

### Finance
- [On-Prem Inbound Automation for Finance](https://clarm.com/blog/articles/on-prem-inbound-automation-for-finance/): Explains on-prem, data residency, SOC 2, and regulated buyer workflows for financial services.

### Category
- [What Is Clarm Atlas?](https://clarm.com/blog/articles/what-is-an-ai-revenue-desk/): Defines the category: capture inbound, qualify buyer intent, and route revenue automatically across channels.
- [AI Lead Qualification vs Traditional Lead Scoring](https://clarm.com/blog/articles/ai-lead-qualification-vs-lead-scoring/): Compares pageview/form scoring with conversation-based qualification and buyer-intent extraction.

### Tactical
- [How to Respond to Inbound Leads Faster Than Your Competitors](https://clarm.com/blog/articles/how-to-respond-to-inbound-leads-faster/): Tactical guide to sub-5-minute response, instant answers, and routing high-intent visitors while motivation is high.

### Decision Guide
- [AI Chatbot vs Live Chat for B2B](https://clarm.com/blog/articles/ai-chatbot-vs-live-chat-for-b2b/): Decision framework for replacing or augmenting human chat with AI-first inbound qualification.

### Pipeline
- [Inbound Pipeline Automation: The Complete 2026 Guide for B2B Teams](https://clarm.com/blog/articles/inbound-pipeline-automation-complete-guide/): Covers capture, qualification, enrichment, routing, measurement, and operational ownership for inbound pipeline.

### Authority
- [The YC Founder's Guide to AI Inbound](https://clarm.com/blog/articles/yc-founder-guide-to-ai-inbound/): Playbook for founders turning launch traffic and early community activity into qualified conversations.

### Compliance
- [SOC 2 Compliant AI Chat for SaaS](https://clarm.com/blog/articles/soc2-compliant-ai-chat-for-saas/): Explains SOC 2 expectations, data controls, no-training commitments, and procurement needs for AI chat.

### Product
- [Slack Sales Alerts for High-Intent Visitors](https://clarm.com/blog/articles/slack-sales-alerts-for-high-intent-visitors/): Explains high-intent visitor alerts, Slack routing, and founder/sales handoff for revenue conversations.
- [Visitor Deanonymization Explained: How to Know Who's on Your Site Before They Fill a Form](https://clarm.com/blog/articles/visitor-deanonymization-explained/): Explains visitor deanonymization, VPN limitations, privacy boundaries, and how intent signals supplement IP matching.

### Vertical
- [AI Inbound for Open Source Companies](https://clarm.com/blog/articles/ai-inbound-for-open-source-companies/): Covers GitHub, Discord, docs, community support, and enterprise buying signals for OSS teams.

## Contact

- Contact page: https://clarm.com/contact/
- Chat: https://clarm.com
- Chat API: POST https://clarm.com/api/agent/chat with {"message":"your question"}
- Email: marcus@clarm.com (sales, press), support@clarm.com (product, privacy)
- Book a demo: https://cal.com/stormm/revenue-desk
- Y Combinator: https://www.ycombinator.com/companies/clarm
- LinkedIn: https://www.linkedin.com/company/clarm
- Twitter/X: https://x.com/clarm_ai
- GitHub: https://github.com/clarm