---
title: "Penetration test results and attestation"
description: "Results of the independent white-box penetration test of Speak AI by Aikido Security, with the attestation letter and remediation summary to download."
---

> Documentation Index
> Fetch the complete documentation index at: https://docs.speakai.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Penetration test results and attestation

Speak AI has completed an independent white-box penetration test by Aikido Security. You can
download the attestation letter and the remediation summary below and pass them to your own
security, procurement or compliance team.

## Who tested Speak AI, and what was in scope?

Aikido Security, an independent security company, ran a white-box web application penetration
test of Speak AI dated 9 October 2026. White box means the testers had full access to the
application and its systems, which makes it the most thorough of the three test types.

The test covered these domains and their endpoints:

- `app.speakai.co`, the Speak AI web app
- `api.speakai.co`, the Speak AI API
- `storage.speakai.co`, file storage

Testing followed the OWASP Testing Guide v4.2, PTES, NIST SP 800-115, OWASP (M)ASVS and the OWASP
AI Testing Guide, including tests of AI agent behavior. It covered the OWASP Top 10 categories,
among them injection, cross-site scripting, server-side request forgery, broken authentication,
insecure direct object references, business logic flaws and exposure of sensitive information.

## What did the test find, and what is still open?

Every critical, high and medium severity finding from the test is closed: each one was fixed, or
reviewed and accepted as a documented risk. As of 19 November 2026, Aikido reports the following
for the tested scope:

| Severity | Open findings |
|---|---|
| Critical | 0 |
| High | 0 |
| Medium | 0 |
| Low | 5 |

The average time to resolve a finding was 38 days. Aikido rates the security posture of the
tested scope **A (Excellent)**, the top of its five-point scale, meaning it aligns with strong
baseline expectations.

## What can I download?

Both documents are published by Aikido Security and shared here unchanged.

- **[Penetration test attestation letter](https://docs.speakai.co/help/security/speak-ai-penetration-test-attestation-2026.pdf)**
  (PDF, 6 pages). Confirms that the test took place, with its scope, methodology and the
  vulnerability classes tested. Aikido issues this letter for sharing with third parties for due
  diligence, RFP responses and compliance.
- **[Penetration test remediation summary](https://docs.speakai.co/help/security/speak-ai-penetration-test-remediation-summary-2026.pdf)**
  (PDF, 5 pages). Shows the findings still open by severity, the average resolution time and the
  security rating after remediation.

## Can I see the full penetration test report?

The full findings report is shared with customers and auditors under a non-disclosure agreement,
because it describes how the platform works internally. To request it, or to talk through the
results with us, [book a security review call](https://speakai.co/book?utm_source=docs&utm_medium=docs&utm_campaign=security).

## How does Speak AI find vulnerabilities between tests?

Speak AI scans continuously between penetration tests. Automated scanning through Aikido Security
covers dependencies, application code and secrets across production. The
[vulnerability management policy](/help/security/policies/vulnerability-management/) describes
how issues are prioritized, fixed and verified.

---
Related: [Security](/help/security/) ·
[Vulnerability management policy](/help/security/policies/vulnerability-management/) ·
[Sub-processors](/help/security/sub-processors/)

Source: https://docs.speakai.co/help/security/penetration-test/index.mdx
