Crogl
Also known as: Crogl Inc, Crogl knowledge engine
Agentic AI security platform running fully autonomous end to end alert investigations with no playbooks and no human trigger, deployed inside the customer environment so data never leaves it, with a free local install tier.
Crogl runs autonomous end to end threat investigations for enterprise security operations. Founded in 2023 by CEO Monzy Merza with CTO David Dorsey and Bradford Lovering, it raised $30 million across a $5 million seed led by Tola Capital and a $25 million Series A led by Menlo Ventures, and launched out of private beta in March 2025 already deployed at Fortune 100 enterprises and government organisations.
The design position is stated bluntly: no playbooks, no schema normalisation, no manual triage, no prompt and no human trigger. An alert arriving from a SIEM, SOAR or ticketing system is detected immediately with no polling delay, and threat intelligence advisories including CRISP reports, ISAC bulletins and vendor advisories are treated as first class inputs alongside alerts rather than as a separate feed. Underneath sits a knowledge graph that continuously maps the environment across users, assets, behaviours, relationships and access patterns, an orchestration layer that plans and executes the investigation itself, and a skills library covering threat hunting, full alert investigation, report generation and incident impact analysis.
The division of labour is explicit and unusually well stated for this category: Crogl handles the investigation and the analyst makes the call. Every action is visible, modifiable and documented, and every investigation arrives fully auditable and decision ready. Given that autonomous investigation is the product, that audit posture is the load bearing claim rather than a compliance footnote.
The deployment model is what separates it from the rest of the AI SOC field. Crogl works against your tools, your workflows and your data, and states that data never leaves your environment. It also ships a free enterprise grade version that installs on a workstation in minutes on macOS or Linux with no procurement cycle, which is a distribution strategy nobody else in this lane runs. Company research it published in 2026 frames the market problem: enterprise teams receive around 4,330 alerts a day and investigate only 37 percent of them.
Vendor details
Canonical URL
https://www.crogl.com
Category
Security / SOC agent
Company status
independent
Use cases & customers
Target customers
Deployment options
In practice
Your team receives thousands of alerts a day and investigates a fraction of them. Crogl investigates every alert autonomously without a playbook and hands back a documented, decision ready case rather than a priority score.
Your data cannot leave your environment for regulatory reasons, which rules out most AI SOC overlays. Crogl runs against your own tools and workflows with data staying in place.
A new ISAC advisory lands. Crogl treats threat intelligence advisories as first class inputs alongside alerts and hunts against them rather than waiting for a matching detection to fire.
Sources & related URLs
Agentic Index coverage score
10.0 / 14 capabilities · 71%
| Integrations & Tool CallingA dedicated tool integration layer queries the existing security stack including SIEM, SOAR and ticketing, with no schema mapping or normalisation required, and the agent uses the customer's own tools during investigation. Crogl product page 2026-08-05 | Full |
|---|---|
| Workflow OrchestrationAn orchestration layer autonomously plans and executes the investigation workflow with no playbook required, gathering context, querying tools and cross referencing across alerts and advisories end to end. Crogl product page and MSSP Alert 2026-08-05 | Full |
| Knowledge Grounding & RAGA knowledge graph continuously maps the environment across users, assets, behaviours, relationships and access patterns, and investigations cross reference that graph against alert and advisory data. The company brands this its knowledge engine. 2026-08-05 | Full |
| Human Oversight & GuardrailsAn unusually explicit division of labour: Crogl handles the investigation and the analyst makes the call, with every action visible, modifiable and documented and investigations delivered decision ready rather than actioned. Crogl homepage 2026-08-05 | Full |
| Security, Identity & GovernanceData never leaving the customer environment is a substantive privacy and compliance control, and the company frames the platform around compliance and reduced operational risk, but no certification list, single sign on or role based access documentation was retrieved. 2026-08-05 | Partial |
| Observability & AuditabilityEvery investigation is fully documented and auditable and every action is visible and modifiable, with generated investigation reports and incident impact analyses. Since autonomous investigation is the product, this audit posture is the load bearing claim rather than a footnote. Crogl homepage and product page 2026-08-05 | Full |
| Memory & State PersistenceThe environment knowledge graph persists and updates continuously and the system is described as learning from analyst feedback over time, but no agent scoped memory or investigation state model is documented as such. 2026-08-05 | Partial |
| Deployment & Data ResidencyRuns against the customer's own tools, workflows and data with the explicit statement that nothing leaves the environment, plus a downloadable install running on a workstation in minutes on macOS or Linux. Rare in a lane where nearly every competitor is a vendor hosted SaaS overlay. 2026-08-05 | Full |
| Prebuilt Agents, Templates & PacksShips a skills library covering threat hunting for indicators of compromise and adversarial behaviour, complete investigation of all alerts, investigation report generation and incident impact analysis. 2026-08-05 | Full |
| Triggers & Channel CoverageAlerts entering from SIEM, SOAR, ticketing or any integrated source are detected immediately with no polling delay and no human trigger, and threat intelligence advisories including CRISP reports, ISAC advisories and vendor bulletins are first class inputs. Crogl product page 2026-08-05 | Full |
| Model Flexibility & RoutingLarge language models are used to weigh evidence and generate findings, but no model list, routing policy, model selection or bring your own key capability is documented. 2026-08-05 | Unable to verify |
| APIs, SDKs & MCP ExtensibilityA tool integration layer and an installable local runtime give real extensibility into the customer stack, but no public API, SDK or MCP surface is documented on the pages retrieved. 2026-08-05 | Partial |
| Testing, Debugging & OptimizationThe system is described as continuously learning from analyst feedback and adapting to new threats, and full auditability makes an investigation inspectable after the fact, but no evaluation harness, published benchmark or regression testing is documented. 2026-08-05 | Partial |
| Browser & Computer UseThe agent queries security tools through an integration layer rather than operating their interfaces, and no browser control or computer use capability is documented. 2026-08-05 | Unable to verify |
The Agentic Index coverage score grades every vendor Full, Partial or Unable to verify against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Pricing
Free enterprise grade tier available for self install; paid terms not published
not disclosed
What is public
The existence and mechanics of the free self install tier are public. No paid rate is.
Billing mechanics
A free enterprise grade version is downloaded and installed by the customer, running in their own environment. Commercial terms above that tier are not published and are negotiated directly.
Cost watchouts
The free tier entered private preview ahead of general availability, so production commitments should confirm current availability and what the paid tier adds
Variable cost rationale
Running inside the customer environment on customer infrastructure moves some cost to the buyer's own compute, and with no published unit for the commercial tier the exposure above the free tier cannot be modelled.
Additional watchouts
Independent buyer guidance in this category warns specifically that per alert pricing ages badly as volume grows; since Crogl has not published its unit, that question should be asked directly. The company is also small and young, founded 2023 with roughly $30 million raised.
Sales call required
Mixed (some tiers require a call)
Free / trial
Free version of the platform for organisations and MSSPs, downloadable and deployable on a workstation in minutes on macOS or Linux, no procurement cycle required
Commercial notes
The free self install tier is a distribution strategy nobody else in the AI SOC lane runs, and it fits the architecture: because the platform executes inside the customer environment against their own tools, there is no vendor hosting cost per trial. That makes free deployment cheap for Crogl in a way it would not be for a SaaS overlay competitor.
Key ambiguities
No paid rate, tier structure or billing unit is published. Whether commercial pricing tracks alert volume, investigations, seats or environment size is unknown, which matters in a category where per alert pricing is a known trap.
Missing data
Paid rate, tier names, billing unit, minimum commitment, and what capabilities the free tier withholds.
Related vendors
- 7AI — Swarming agentic SOC from the Cybereason founders: sixty plus domain…
- Abnormal AI — Behavioural AI email security with three named autonomous agents: a…
- AirMDR — AI-native MDR built around a Virtual Analyst that triages 95% of…
- Airrived — Agentic OS that unifies SOC, GRC, IAM, vulnerability management, IT,…
- Andesite — Bionic SOC: a human-AI collaboration platform that automates triage,…
- Anvilogic — Agentic security operations platform that decouples detection from…
Alternatives to Crogl
The closest documented capability profiles to Crogl among security and SOC agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Kai10.5 / 14Fuller documented coverage on Security, Identity & Governance
- Abnormal AI9.0 / 14A lighter documented profile than Crogl
- Arctic Wolf9.0 / 14A lighter documented profile than Crogl
- D3 Security (Morpheus)10.0 / 14Fuller documented coverage on Security, Identity & Governance
- Darktrace9.0 / 14A lighter documented profile than Crogl
- Seemplicity10.0 / 14Fuller documented coverage on Security, Identity & Governance
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded