Block Coplan
https://github.com/block/coplan- Category
- AI Agents
- Rank
- No. 1474Tools index
Previous survey · No. 1480 ·
- Pricing
- Open Source
- Type
- TOOL
- Builder
- block
- GitHub
- 32 stars
- Latest release
- v1.0.0
- Date
About
Planning with agents — a collaborative planning tool for AI coding workflows.
What it does
CoPlan hosts Markdown engineering plans for team review before implementation starts. Reviewers select passages and leave threaded comments. Accepted feedback lets an AI agent make constrained text edits, while the system preserves immutable versions, change reasons, model details, prompts, and the comments behind each revision.
Why it's ranked here
The strongest idea is its deliberate split between human judgment and agent editing. Inline discussion stays familiar, but leases, base revisions, deterministic operations, and immutable history put useful controls around automated changes. That combination makes it credible for teams that need agent assistance without losing accountability.
What's good
Semantic edits target exact text, headings, or matching paragraphs instead of fragile line numbers. Ambiguous matches fail rather than guessing. A single edit lease prevents competing writers, and every resulting version records provenance. Realtime updates, lifecycle states, private brainstorms, organization publishing, attachments, and optional automated reviewers support an actual review process.
Tradeoffs
The asymmetric model means humans comment while agents perform edits, which will not suit teams expecting direct collaborative authoring. Deployment requires a Rails host, database migrations, reference seeds, authentication wiring, background jobs, and realtime infrastructure. Existing Active Storage installations using integer polymorphic identifiers need a schema change before attachments work.
How to use it well
Use CoPlan when a Rails-based engineering organization drafts plans in Markdown, wants domain experts commenting inline, and can let agents apply accepted changes through the API. It fits between initial drafting and implementation, especially where auditability matters. It does not cover code execution or implementation; its stated purpose ends before coding begins.
Technical notes+
README.md presents a Rails 8 engine installed as coplan-engine, with bearer-authenticated REST endpoints for reading plans, acquiring leases, applying operations, and posting comments. docs/HOST_APP_GUIDE.md requires mounting CoPlan::Engine, installing migrations, running coplan:seed, and configuring CoPlan.configure with an authentication callback. docs/PLAN.md specifies deterministic replace_exact, insert_under_heading, and delete_paragraph_containing operations against a base revision. Gemfile includes Rails, MySQL and PostgreSQL adapters, Hotwire, SolidQueue, SolidCable, CommonMarker, Diffy, and ruby-openai. config/ci.rb runs style, dependency audits, Brakeman, Rails tests, system tests, and seed tests. db/schema.rb shows UUID-backed CoPlan records, edit leases, edit sessions, comments, versions, libraries, folders, events, and API-token digests.
Observed
- License
- Apache License, Version 2.0
- Primary language
- Ruby, implemented with Ruby on Rails 8+
- Packaging
- Rails engine distributed through the coplan-engine gem
- Install surface
- Gemfile dependency, mounted Rails engine, migrations, and idempotent reference-data seed
- Interfaces
- Web interface and bearer-authenticated REST API; optional Slack Events integration
- Database support
- MySQL and PostgreSQL adapters are included for host applications
- Realtime and jobs
- Turbo Streams with SolidCable for realtime updates and SolidQueue for background jobs
- Verification
- CI defines Ruby style checks, dependency audits, Brakeman analysis, Rails tests, system tests, and seed tests
Read from README.md, Gemfile, docs/PLAN.md, docs/HOST_APP_GUIDE.md, db/seeds.rb, config/ci.rb, db/schema.rb, config/boot.rb, config/puma.rb, config/routes.rb, db/cable_schema.rb, db/cache_schema.rb, db/queue_schema.rb, config/importmap.rb, config/application.rb.
What it can do
Create collaborative AI coding workflows
Project requirements and team specifications → Structured coding workflow plan
Coordinate tasks between AI agents and human developers
Development tasks and agent capabilities → Task assignments and coordination timeline
Generate code using AI agents
Natural language requirements and specifications → Source code files
Track progress of multi-agent coding projects
Agent activities and task completion status → Project progress reports and status updates
Plan development sprints with AI assistance
Project goals and timeline constraints → Sprint plans with task breakdowns
Manage collaboration between multiple AI coding agents
Agent configurations and project scope → Coordinated multi-agent workflow execution
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.