Roadmap

Requirements as files in git

One Markdown file per element, nested by parent, with status, relationships and acceptance criteria. The model describes the system as it is now; history stays in git and in the journal.

Acceptance criteria inside the requirement

A criterion is a command with an expected result, or a check by a judge with the evidence named. spec-editor validate checks the model and its links to code.

Methodology as data

Aspects, element types and relationships are declared in a YAML file. The bundled methodology is translated into five languages, and a methodology can be generated for a domain.

Sources from your channels

Documents, bugs, incidents and ideas arrive as sources. Channels are Jira, GitHub Issues, Slack, Discord and Telegram, plus import of OpenSLO files.

CLI and MCP server

A command line and an MCP server for external agents, so they receive the elements they need instead of a whole document. Automatic set-up for Claude Code, Claude Desktop, Codex CLI, Gemini CLI, Cursor, VS Code, Windsurf, Zed, OpenCode and JetBrains IDEs.

Editor extensions

Extensions for VS Code, JetBrains IDEs and Zed.

Several projects in one installation

Several independent projects run from a single installation.

Export

SRS in the style of IEEE 830, styled HTML and TRLC.

Web interface

Element tree, element card, trace map, interactive graph editor with several diagram types, run log and run-time trace map. Light and dark themes.

Run-time trace and hierarchical logging

Code marked with an element writes operation events. The log follows the tree of requirements, and the run-time trace map shows which requirements actually ran.

Model providers

Your own Claude Code subscription (no API key), API keys through LiteLLM, and local models through Ollama.

Delivery team of agents (Pro, early access)

Analysts, an analysis manager, a project manager, a coder, a tester and devops, with review stages decided by evidence. Pro is in early access; purchase is not open.

Test stands, rollback and cost control (Pro, early access)

Every plan runs its acceptance criteria on a blue/green test stand before promotion. Model ladder, budget guards and a pool of subscription accounts with failover.

Breakpoints (Pro, early access)

Visual comments right in your running product. A click on an element of the screen creates a source bound to the requirement of that element.

Reengineering and coverage (Pro, early access)

Requirements restored from your existing codebase, and spec-editor coverage showing coverage of requirements by code, both ways.

Monitoring (Pro, early access)

Repeating errors in the logs are recognised by signature and become incidents. An on-duty role sorts them and sends them back as sources.

Self-improvement (Pro, early access)

KPIs of the model are evaluated from events. An observer proposes hypotheses, experiments decide, and roles learn from finished runs.

Automatic policies and a fast path (Pro, early access)

Routine operator decisions are taken by policy, and one-line edits take a short route.

User manual and demo video (Pro, early access)

Both are produced from the user scenarios of the model. The video is recorded on a stand.

Self-service purchase and licence keys

Buying a licence without contacting us, with the Pro package as a download.

Account area on the site

Seats and subscriptions in one place.

Hosted view of a model

coming soon

Look at a model in a browser without running anything locally.

Seats for a studio's clients

coming soon

Give a studio's clients their own seats.

Projects of one organisation linked together

later

A federation connector that links the separate projects of one organisation. It will be a separate paid product, sold apart from spec-editor Pro.