Vecbase

Vecbase is an AI-native company operating system that lets teams build and run a workforce of specialized AI agents inside a single shared workspace. Each agent has its own role, persistent memory, file access, and a secure sandbox, and can be assigned tasks that run autonomously on schedule or on demand. The platform is available as a web app and a downloadable desktop app.

What It Is

Vecbase positions itself as a "one workspace, one AI team" platform where AI agents function like employees — each with a defined job (Researcher, Coder, Writer, Operator), access to shared files, and the ability to use real tools such as web search, code execution, and third-party app integrations. Rather than a single chatbot, the product is designed around parallel, role-specific agents that hand off work to each other and report results back to the user.

How the Agent Workspace Works

The core unit is the workspace, which contains a team of agents, a shared Drive, and a set of connected tools and services. Key structural elements include:

• Agents: Each agent has its own settings, file space, skills, and schedules. Multiple agents run in parallel without interfering with each other.
• Drive: A shared file library where agents read and write files — PDFs, spreadsheets, markdown docs, code — accessible to every agent in the workspace.
• Skills: Reusable workflow routines that can be written once and executed by any agent, eliminating repeated instructions.
• Schedules: Agents can be set to run on a cron-style schedule (e.g., daily 9 AM reports, weekly Monday briefs, six-hourly checks) without manual triggering.

Tools, Capabilities, and Integrations

Every agent action — web searches, document reads, code edits, shell commands — streams into the chat in real time and is replayable. Vecbase bundles managed capabilities from multiple providers under a single bill:

• Speech recognition: ElevenLabs Scribe
• Text-to-speech: ElevenLabs, Google Gemini
• Document parsing: Reducto, Unstructured
• Image generation: OpenAI, Google Gemini
• Web search / deep research: Perplexity, Exa, Google Gemini
• Web crawling: Cloudflare Browser Rendering

App connectors (authorized once via OAuth) include Gmail, Google Calendar, Google Drive, Slack, Discord, Notion, GitHub, Linear, Jira, Asana, Figma, and HubSpot.

Model Choice and Billing Transparency

Vecbase supports multiple LLM providers per agent, allowing each agent to use the model best suited to its task. Supported models include Vecbase's own Vecbase 1.0 family (Lite, standard, Pro), OpenAI GPT, Anthropic Claude, Google Gemini, DeepSeek, Kimi/Moonshot, MiniMax, and Qwen. The platform advertises transparent per-token billing with no model lock-in — all usage is deducted from a monthly credit balance included with each subscription, with overages billed at the same published rates.

Deployment and Platform

Vecbase runs as a web application and also offers a downloadable desktop app. The platform is multilingual, with the interface available in English, Simplified Chinese, Japanese, Korean, French, German, Spanish, Portuguese, Russian, Arabic, Hindi, and Italian — suggesting a global audience. Documentation is hosted at docs.vecbase.com.