Specivo Docs

What is Specivo?

Specivo is a self-hosted platform that combines four things most teams otherwise stitch together, or don't have at all:

  1. A project tracker — issues, sprints, releases, and reporting.
  2. A wiki — your team's living knowledge base, version-controlled.
  3. An AI integration layer — a built-in MCP server so AI assistants can do real work in the tracker.
  4. AI usage and spend tracking — what each person, project and feature costs in AI, tied to the issues the work was done for.

Because you run it yourself, your issues, documents, and search index never leave infrastructure you control. There is no per-seat license; you add as many teammates as you need.

Know what your AI work costs

Your team uses Claude Code, Codex, Cursor and other AI tools every day, on personal plans, company seats and API keys. The invoice says how much was spent. Specivo shows who spent it, and on which feature. Most trackers can't answer that, because they never see the AI work at all.

  • Collected automatically. The open-source ai-usage reporter reads the local logs of Claude Code, Codex, Gemini CLI and OpenCode and sends only usage numbers: never prompts, code or file contents. See Reporting AI usage.
  • Tied to the issue. Each session's usage is attributed to the issue the agent worked on, so every issue shows what it cost in AI, its sub-issues included. See AI usage.
  • Every kind of cost in one view. Personal plans, team seats and pay-per-use API keys, in any currency, priced from a bundled list of models.
  • Drill down. From the company total to a project, a person, a feature and the sessions behind it. See AI spend reports.

The AI usage tab of an issue: what the feature cost, by person, tool and model

Who it's for

Specivo is built for teams who want one place for what we're doing and what we know, and who increasingly work alongside AI agents. If your decisions vanish into chat, your docs are scattered across five apps, and your AI assistants ship work in a black box at a cost nobody can break down, Specivo is designed to fix that.

How the pieces fit together

Everything in Specivo connects:

  • An issue can link to related issues, carry structured metadata, hold file attachments, and reference a wiki page.
  • A wiki page can cross-link to other pages and be referenced from issues.
  • Search spans issues, wiki pages, comments, and attachment descriptions at once.
  • AI agents reach all of the above through the MCP server, and every change they make is recorded just like a change made by a person.
  • AI usage is recorded against those same issues and projects, so the cost of a feature sits next to the work itself.

AI is optional

Specivo works fully without any AI features. Semantic search and the MCP server are there when you want them — using a bundled on-device model or your own API key — and out of the way when you don't.

What this guide covers

This is the end-user and operator guide. It explains how to use Specivo (issues, metadata, wiki, AI agents) and how to install and run it (Installing Specivo). It is not an API or source-code reference.

Ready? Continue to Core concepts, or jump straight into Your first 10 minutes.