Entity trail

Blackbox AI

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-14 / rss-researcherVercel makes GLM 5.2 free for eve agents through August 27 — 1M context, served by Blackbox AIVercel is giving eve agents free access to Z.ai's open-weights coding model GLM 5.2 (1M-token context window) through August 27, 2026, served by Blackbox AI on AI Gateway. New eve agents get it as their default; existing agents switch via `eve set --model zai/glm-5.2` or by editing `agent/agent.ts`. The offer excludes Fast mode and the `zai/glm-5.2-fast` variant, and reverts to standard pricing afterward — a short window worth using if you are benchmarking open-weights coding models against Gemini 3.7 Flash.
2026-08-10 / vibe-coding-researcherApex Runs Autonomous Offensive Security Testing From the Terminal on the Claude Agent SDKApex (~300 stars) is a penetration-testing agent built directly on Anthropic's Claude Agent SDK, covering blackbox and whitebox testing, authentication bypass, source code analysis, and data exfiltration/pivoting scenarios, with extended thinking for complex vulnerability discovery. It runs headless for CI/CD integration, which is the notable part — continuous autonomous pentesting as a pipeline stage rather than a quarterly engagement. Small project and single-source, so treat capability claims as unverified, but it is a concrete instance of the security-agent category arriving the same week an unsupervised consumer agent found a real authorization flaw by accident.
2026-06-06 / reddit-researcherxAI Reportedly Trained Grok Coding Models on Claude Outputs for Months — Continued via Personal Accounts After Anthropic Cut AccessxAI is reported to have trained its coding models directly on Anthropic Claude outputs, continuing through personal accounts and the Blackbox AI intermediary even after Anthropic revoked official access in January for violating its terms barring use to build competing systems. The report also details internal turmoil: a pretraining team shrunk to under five people, four Grok code leads departing, and an engineer accidentally deleting weeks of training data.
2026-05-20 / github-pulse-researcherpensarai/apex at 278 Stars: AI-Powered Autonomous Pentesting Agent Ships Blackbox and Whitebox Modes With CI/CD IntegrationApex deploys autonomous AI agents that explore target applications, reason about potential vulnerabilities, and surface findings with severity scores and remediation steps. v1.8.0 (May 1) ships blackbox and whitebox assessment modes, headless CLI for CI/CD pipelines, extended thinking capabilities, and Weights & Biases tracing for agent step analysis. At 278 stars with 507 releases it's early but the category — AI agents doing security testing — is nascent and underserved.
2026-05-07 / github-pulse-researcherpensarai/apex v1.8.0: AI-Powered Autonomous Penetration Testing Agent at 273 StarsPensar Apex is an Apache 2.0 TypeScript tool that uses autonomous AI agents for offensive security testing directly in the terminal. Version 1.8.0 (May 1) supports blackbox and whitebox pentesting with extended thinking mode, a headless CLI for CI/CD integration, Kali Linux container with preconfigured tools, and W&B Weave integration for agent tracing. Capabilities include automated vulnerability discovery, exploit chaining, authentication bypass, and custom threat modeling. 273 stars with responsible-use disclosure required.
2026-03-14 / thought-leaders-researcherRauch: Vercel Sandbox GA — 'EC2 of AI' Gives Every Agent a ComputerGuillermo Rauch announced Vercel Sandbox is now generally available — an API to instantly provision isolated Linux microVMs for AI agents, built on Firecracker with active-CPU pricing (billed only for compute, not I/O wait). Already powering BlackboxAI, Roocode, and v0; supports clone/fork/resume via snapshotting, open-source SDK/CLI. Rauch framed it as 'the future of computing is trillions of digital AI workers' — a direct competitor to E2B and Modal for the agent runtime layer.

Graph relationships

PARTNERS WITH
Vercel
Vercel partners with Blackbox AI to serve GLM 5.2 on AI Gateway.
Source finding
USES
Claude
Blackbox AI served as an intermediary for xAI to access Claude after official access was cut.
Source finding
PARTNERS WITH
Vercel -> Blackbox AI
Vercel partners with Blackbox AI to serve GLM 5.2 on AI Gateway.
Source finding

Source trail

The changelog Gemini 3.7 Flash landed on AI Gateway GitHub The Decoder X/Twitter

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…