Open-source CLI coding agent that turns ChatGPT Web into Codex and Claude Web into Claude Code. Reads local files, edits code, runs commands and tests in agentic loops with your existing web AI account - no model API key. Also supports DeepSeek, Gemini, Grok, Kimi and GLM. Install with npm i -g wtagent.
Code Generation
other
Approved & Active
0 defined
Sep 2026
Available
UncensoredLabs provides uncensored and abliterated text language models through one [OI]-compatible API. Developers and agent builders use it as the inference backend for chat, coding and automation workloads that need text generation without refusal behaviour on legitimate prompts. The API supports streaming chat completions, prepaid token-based billing, per-API-key spending limits, model allowlists and usage tracking, and works with existing [OI]-compatible clients by changing only the base URL. Eligible new accounts can receive a one-time $0.10 trial usage credit; verification may be required. Ordinary usage is prepaid, with no recurring free tier.
The Hive Code is the home of hive coding: software built by communities and shipped by AI agents. Vibe coding is one person and one AI. Hive coding is many people and many AIs. Anyone can post an idea for software they wish existed, or back an idea that is already moving. The community comments and votes, builders raise their hands, and a person reviews and approves every change. AI agent builds come next. It is free, no code or GitHub is needed to post an idea, and you can look around without an account. It just opened. Founded by Grub Bussio, who has built and sold two e-commerce companies with eight figures in e-commerce sales.
Zozo HQ orchestrates coding agents from dependency-ready issues to reviewable pull requests. It gives each attempt an isolated worktree, pauses for human clarification when needed, and keeps an append-only execution journal for review and recovery. For developers who already use Claude Code, Codex, or OpenCode with OpenRouter model access. OpenCode is the runtime; OpenRouter supplies model access. The required runtime must be available in the deployed Stack image. Model usage is supplied and billed separately from Zozo. Supported workflows: GitHub repositories with GitHub Issues, or matched Jira Cloud + Bitbucket Cloud projects. GitHub + Jira and Atlassian Server/Data Center are not supported. A Stack can contain both GitHub and Atlassian projects, but each project keeps its matched repository/tracker pairing. Atlassian prototyping is not supported; Medium/Large Intake and Wayfinder require a Spec Issue Type. Free Local Stack on your own Mac or Linux; paid managed hosting from $40 USD/month before tax. The local stack runs as you and stops with its terminal. This is an orchestration layer around coding runtimes, not a proprietary foundation model. Docs: https://docs.zozohq.com/guide/ Atlassian announcement: https://zozohq.com/announcements/atlassian-projects OpenCode requirements: https://docs.zozohq.com/guide/task-connect-opencode Disclosure: Submitted by Ryan Graham, Zozo founder, with AI assistance.
Seedance 2.5 creates controllable AI videos up to 30 seconds long in a single pass. It turns text, images, video and audio into finished clips, so creators no longer need to stitch many short generations together by hand. Output renders in up to 1080p, ready for social platforms, ads and presentations. Key capabilities include text-to-video and image-to-video modes, up to 50 multimodal references for character, motion and style consistency, single-pass clips up to 30 seconds, and flexible editing to refine a result instead of starting over. It is suited for marketing demos, anime-style scenes, short-form social content and cinematic concepts. Marketers, filmmakers, animators and independent creators use Seedance 2.5 when they need longer, more consistent AI video without heavy post-assembly. Subscription-based paid plans are available on the website.
Content CreationFluxtail gives engineering teams one place to collect, search and investigate application logs. Its hosted Model Context Protocol (MCP) service lets compatible AI tools search logs and work with streams, receivers and sources. Teams can also use the browser interface for filtered log searches and Live Tail. Shared receivers accept HTTP JSON, OpenTelemetry over HTTP or gRPC, and Prometheus traffic. Paid plans add dedicated receivers for protocols such as Syslog and GELF, plus native sources including Kafka, MQTT and S3/ALB. Accounts keep their logs isolated, and account-scoped access controls determine what users and connected tools can read or change. Fluxtail is useful when investigating application errors, following activity across services, or bringing log evidence into an AI-assisted development workflow. The web interface remains available for teams that prefer to investigate directly in the browser. New accounts can start with a 14-day trial without a credit card. The trial excludes dedicated ports and native sources and expires into read-only access. Starter and Pro are the paid plans; current prices and limits are published at https://fluxtail.io/pricing. Setup guides and protocol documentation are available at https://docs.fluxtail.io/.
DevOpsSandBase gives developers one account and one API key for models, data APIs, sandboxes, and agents — OpenAI-compatible Agents API, E2B-compatible sandboxes, and unified billing.
DevOpsGet notified when we add more code generation agents or agents similar to WTAgent.
List your AI agent in the directory and reach thousands of developers.