How much does Repello AI cost?

No published prices; demo on request. Open-source tools (Whistleblower, Agent-Wiz, the Artemis GitHub Action) are free.

What are the best alternatives to Repello AI?

Based on documented capability coverage across the Agentic Index 14 point taxonomy, the closest alternatives to Repello AI among security and SOC agents are Lasso Security, Noma Security, Straiker, Terra Security. Each is graded on the same public evidence, and no vendor pays for placement.

Does Repello AI support MCP and tool calling?

Full under the Agentic Index verification standard. Repello AI publishes public evidence of Model Context Protocol (MCP) support, tool calling, and API or SDK extensibility for connecting agents to external tools. Evidence basis: Repello publishes a Python client SDK for triggering actions on the Artemis platform (repello-artemis-sdk), an Artemis GitHub Action and a Claude Code plugin that connects to the hosted Artemis Recon server. Its open source tools also inclu. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Does Repello AI give you tracing and observability into agent runs?

Full under the Agentic Index verification standard. Repello AI publishes public evidence of tracing, observability and audit logging into agent runs, so a team can see what an agent did and why. Evidence basis: ARGUS captures every call and anomaly in an activity center with full logging, audit trails and alerts for the customer's AI applications, recording each call, and ARTEMIS reports show attack paths from goals to prompts with evidence. Works. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Does Repello AI have SOC 2, SSO and role based access control?

Partial under the Agentic Index verification standard. Repello AI publishes some public evidence of security and identity controls, but narrower than the full set of SOC 2, single sign on (SSO) and role based access control (RBAC). Evidence basis: Repello carries AICPA SOC 2 and ISO 27001 marks, but it names no SSO, role model or audit surface for its own console. Workstation Lens policies are managed centrally, with audit trails, for the customer's coding agents. Repello AI, repello. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Back to vendors

Repello AI

Also known as: Repello, ARTEMIS

Visit site

Security / SOC agentindependentVerified 2026-09-28

Entry priceContact sales; no prices published. Open-source tools are free.Full pricing detail

Autonomous red teaming of the customer's AI apps and agents (ARTEMIS), runtime guardrails that log every call (ARGUS), endpoint agent discovery (Workstation Lens) and an AI inventory, with a Python SDK and GitHub Action.

Repello AI sells four products for securing the AI an enterprise builds and uses. ARTEMIS is an autonomous red teamer. Its reasoning adversaries run parallel attack sequences over many turns against the customer's AI applications and agents through APIs and the browser, chain exploits across integrations, and report exploit chains of several stages with risk scores, evidence and mitigations mapped to MITRE, OWASP and NIST, to ISO 42001 and to the EU AI Act, drawing on a library of 15M+ curated attack vectors across text, image and audio.

ARGUS is a runtime guardrail that enforces custom, context aware rules, blocks unsafe actions and logs every call and anomaly in an activity center with audit trails. Workstation Lens finds unmanaged coding and browser agents such as Claude Code, Cursor, Codex and Copilot on employee devices, detects credential access, tracks MCP servers and deploys fixes through EDR and MDM tools. INVENTORY maps AI applications, workflows and exposure paths. Repello publishes a Python SDK and GitHub Action for Artemis and open source tools such as Whistleblower and Agent-Wiz, and shows SOC 2 and ISO 27001 marks.

Vendor details

Canonical URL

https://repello.ai

Funding status

Independent AI security company; funding amounts are not disclosed. Customer red-teaming engagements include enterprises such as Lyzr, and Repello maintains public open-source tooling on GitHub (Artemis GitHub Action, Whistleblower system-prompt-leakage tester, awesome-llm-redteaming).

Company status

independent

Use cases & customers

Primary use cases

autonomous AI red teaming of LLMs, agents, and MCPmulti-stage exploit chain discovery and attack-path visualizationAI asset discovery and AI Bill of Materials (AI BOM)adaptive runtime guardrails calibrated from red-teaming resultsCI/CD-integrated continuous AI security with compliance-mapped reporting

Target customers

enterprises deploying LLMs, AI agents, and MCP in productionAI product and engineering teams needing pre-release security testingsecurity teams demonstrating compliance coverage for AI systemsorganizations securing multi-modal AI applications

Deployment options

SaaS; hosting environment not stated on the pages readCI/CD through the Artemis GitHub Action

Integrations

ARTEMIS tests through API endpoints and the browser and runs in CI/CD via a GitHub Action and Python SDK. Workstation Lens covers Claude Code, Cursor, Codex, Copilot and browser agents and deploys fixes through CrowdStrike, SentinelOne, Defender, JAMF, Kandji and JumpCloud. A Claude Code plugin connects to the hosted Artemis Recon server.

In practice

You are about to ship an AI agent with tool access. Repello's ARTEMIS red teams it through the API and the browser, chains exploits across its integrations and hands back risk scores, evidence and fixes mapped to OWASP and NIST.

Your production chatbot needs guardrails that stop abuse as it happens. ARGUS enforces your custom rules on every call and logs each anomaly for audit.

Engineers have installed coding agents and MCP servers on their laptops. Workstation Lens finds them, blocks unsafe actions under your policies and removes risky plugins through CrowdStrike, SentinelOne or JAMF.

Sources & related URLs

Research sources

https://repello.ai/ https://repello.ai/product https://repello.ai/blog/introducing-artemis-automated-red-teaming-to-secure-your-ai-applications https://github.com/Repello-AI

Agentic Index coverage score

8.0 / 14 capabilities · 57%

Integrations & Tool Calling	Full
ARTEMIS attacks the customer's AI through API endpoints and the browser and runs in CI/CD, and Workstation Lens deploys fixes for endpoint agents through CrowdStrike, SentinelOne, Defender, JAMF and Kandji. JumpCloud is now among those channels, and Workstation Lens can remove, update, disable or restrict a risky MCP server, skill or plugin on every affected employee device. SourceRepello AI, repello.ai/product and /workstation-agent; repello.airead 2026-10-05
Workflow Orchestration	Full
ARTEMIS deploys reasoning adversaries that orchestrate sophisticated attack sequences over many turns. It runs parallel attack threads and chains exploits across integrations, adapting its tactics to each response, so several of Repello's own agents attack at once. SourceRepello AI, repello.ai/productread 2026-10-05
Knowledge Grounding & RAG	Partial
INVENTORY maps the customer's AI applications, workflows and exposure paths, and attacks draw on Repello's own library of 15M+ curated attack vectors. There is no retrieval over the customer's own material. INVENTORY builds a living AI bill of materials of every AI system in production, and Workstation Lens finds AI agents and their MCP servers, skills and plugins across employee devices. SourceRepello AI, repello.ai and /workstation-agentread 2026-10-05
Human Oversight & Guardrails	Partial
ARGUS enforces custom, context aware rules that the customer sets across modalities, languages and domains, and blocks unsafe actions at runtime. Workstation Lens checks each coding agent action against the customer's policies and blocks unsafe ones before they run. No step has a person approve or get asked before an agent action goes through. SourceRepello AI, repello.ai/argus and /workstation-agentread 2026-10-05
Security, Identity & Governance	Partial
Repello carries AICPA SOC 2 and ISO 27001 marks, but it names no SSO, role model or audit surface for its own console. Workstation Lens policies are managed centrally, with audit trails, for the customer's coding agents. SourceRepello AI, repello.ai and /workstation-agentread 2026-10-05
Observability & Auditability	Full
ARGUS captures every call and anomaly in an activity center with full logging, audit trails and alerts for the customer's AI applications, recording each call, and ARTEMIS reports show attack paths from goals to prompts with evidence. Workstation Lens follows prompts, tool calls, commands, file and credential access and outbound destinations as coding agents work, with audit trails. SourceRepello AI, repello.ai/argus, /product and /workstation-agentread 2026-10-05
Memory & State Persistence	Not documented
The agents keep no memory. The INVENTORY map covers the customer's AI estate, and ARGUS logs calls for audit. SourceRepello AI, repello.ai and /argusread 2026-10-05
Deployment & Data Residency	Not documented
ARTEMIS and ARGUS have no named hosting region, customer environment or data residency option, and Repello does not say where test data is kept. ARTEMIS tests as a black box, and ARGUS works as production guardrails. Workstation Lens rolls out to employee devices through the customer's existing endpoint tools. SourceRepello AI, repello.ai/argus and /workstation-agentread 2026-10-05
Prebuilt Agents, Templates & Packs	Partial
Repello ships a library of 15M+ curated attack vectors and prebuilt compliance mappings to MITRE, OWASP and NIST, to ISO 42001 and to the EU AI Act. The catalog is not named, and there are no prebuilt agents. Repello's products split by job. ARTEMIS handles red teaming, ARGUS runtime guardrails, INVENTORY discovery of AI systems and Workstation Lens coding agents on employee devices. SourceRepello AI, repello.ai/product and repello.airead 2026-10-05
Triggers & Channel Coverage	Full
ARTEMIS runs in CI/CD through a GitHub Action, and ARGUS acts on every call in real time, so neither waits for a person to start it. Workstation Lens also checks each coding agent action against policy as it happens, and Repello says ARTEMIS finds the first AI vulnerability in under a minute. SourceRepello AI, repello.ai/product and /workstation-agentread 2026-10-05
Model Flexibility & Routing	Not documented
The customer's models are what Repello tests and guards. Repello offers no model choice or routing of its own. Its testing agents work as reasoning adversaries, and Repello names no model provider behind them. SourceRepello AI, repello.ai and /productread 2026-10-05
APIs, SDKs & MCP Extensibility	Full
Repello publishes a Python client SDK for triggering actions on the Artemis platform (repello-artemis-sdk), an Artemis GitHub Action and a Claude Code plugin that connects to the hosted Artemis Recon server. Its open source tools also include Agent-Wiz, a command line tool for threat modeling agents built with LangGraph, AutoGen and CrewAI, and a free red teaming scan runs at recon.repello.ai. SourceRepello AI, github.com/Repello-AI and repello.airead 2026-10-05
Testing, Debugging & Optimization	Full
ARTEMIS autonomously red teams the customer's AI applications and agents through API and browser, finding and validating exploit chains of several stages with risk scores, evidence and mitigations mapped to compliance frameworks, and runs in CI/CD. It profiles each system's capabilities to tailor its attacks, covers all ten OWASP Top 10 for LLM categories, and tests text, image and audio in 100+ languages. SourceRepello AI, repello.ai/productread 2026-10-05
Browser & Computer Use	Not documented
Customers get no browser automation from Repello. ARTEMIS combines browser and API red teaming to test the customer's applications. Workstation Lens watches coding agents such as Claude Code, Cursor, GitHub Copilot, ChatGPT Desktop, Windsurf, OpenCode, Ollama and Codex on employee devices and does not operate a computer itself. SourceRepello AI, repello.ai/product and /workstation-agentread 2026-10-05

The Agentic Index coverage score grades every vendor Full, Partial or Not documented 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

Contact sales; no prices published. Open-source tools are free.

Inference, not stated by the vendor: AI systems under red teaming and runtime protection, and endpoints covered.

What is public

Product names and capabilities and free open-source tools are public; no plans or prices are disclosed.

Variable cost rationale

Inference, not stated by the vendor: cost likely grows with the AI systems and endpoints covered.

Sales call required

Yes, required for paid access

Free / trial

Free open-source tools (Whistleblower, Agent-Wiz, Artemis GitHub Action); platform demo on request.

Agentic Index verified 2026-09-28

Related vendors

Alternatives to Repello AI

The closest documented capability profiles to Repello AI 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.

Lasso Security7.0 / 14A lighter documented profile than Repello AI
Noma Security7.0 / 14A lighter documented profile than Repello AI
Straiker8.5 / 14Adds documented Deployment & Data Residency
Terra Security6.5 / 14A lighter documented profile than Repello AI
Capsule Security9.0 / 14Fuller documented coverage on Knowledge Grounding & RAG and Human Oversight & Guardrails
Equixly6.0 / 14A lighter documented profile than Repello AI

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Compare Repello AI with other vendors →Compare all security and SOC agents →See where Repello AI ranks →

Data confidence: medium

Head to head

Repello AI vs DeepKeepRepello AI vs Mindgard