AI project status dashboard
Never open a cold project again.
sitrep is a state layer on top of the markdown your AI agents already generate. It reads the plans, decisions and progress your agents already write - no new format, no import - and answers, at a glance and across every project: what moved, what is stalled, and what is waiting on you.
In development. Launching first on macOS. Local-first - your files never leave your machine.
The first fifteen minutes of every session
Run Claude Code, Cursor or Codex across a few projects and you accumulate plans, analyses and decision documents you never asked for and cannot keep in your head. The state of your work is written down - it is just scattered across hundreds of markdown files in a dozen repos.
So every return to a cold project starts the same way: scrolling old chats, re-opening docs, re-deriving where things stand. sitrep exists so that re-entry stops costing the first fifteen minutes of a session. Open it, see the situation, resume.
How it works
- 01
Point it at folders
No import, no vault. sitrep indexes the markdown already on your disk and watches it with kernel events - a corpus of tens of thousands of files narrows to the few hundred documents that are actually project docs.
- 02
It extracts the state
Status lines, phase headings, decision tables, due dates - the conventions agents and humans already use are parsed into structured state. Extraction vocabularies are editable settings, never demands on your files.
- 03
You triage, not archaeology
An email-client-style overview: newest activity first, pending decisions surfaced, what the docs claim beside when they last changed. Click through to read any doc in full fidelity.
The product contract
Eight promises every feature is tested against. A feature that breaks one is wrong even if it is useful.
Zero adoption cost
sitrep reads what agents already write. No new format, no frontmatter, no import, no restructuring. Point it at folders.
State first
Status, phases, decisions and staleness, glanceable across projects - an email client for project state, not a file tree.
Reading parity
Full GFM, highlighted code, Mermaid, KaTeX, callouts. Rendering is the floor, not the product - and never the excuse.
Nothing takes the page away
Live refresh lands in place, scroll re-anchors, reload restores your position. Esc peels one layer at a time.
One writing surface
Three convention-shaped ops - status, due, supersede - behind a preview and a content-hash guard. No other code path ever writes to your files.
Local-first and private
Fully functional forever with no account and no network. No telemetry by default. Your files never leave your machine.
Fast
Instant startup on kernel file events - never a tree walk. Instant search. Work stays proportional to what changed, never to corpus size.
Honest
Claims are measured before they are made - in the UI, in comparisons, and in community participation. Concrete and checkable beats superlatives.
What sitrep will never be
Not a viewer, not an editor, not a Notion/Obsidian replacement. The anti-roadmap is written down and binding - sitrep will never be:
- a general markdown editor, or one more pretty viewer
- an agent orchestrator or parallel-session runner
- a task manager with a format of its own
- a second brain, vault or knowledge base
- live agent control from your phone
- a cloud-only or account-required product
Where sitrep fits
Every adjacent tool gets project state another way: you author it in the tool's shape, you work inside the tool, it reads the session trail instead, you ask an agent to re-derive it every time - or it stays in your head. sitrep parses the state your files already contain. Each comparison below is honest about when the other approach is the better choice.
| Tool | What it is | How you get project state | Comparison |
|---|---|---|---|
| sitrep | AI project status dashboard | Point it at folders - it parses the markdown your agents already wrote, across every project | - |
| Backlog.md | Git-native task board | You and your agents author tasks in its format, per repo | vs Backlog.md |
| Obsidian + Dataview | Knowledge base + query plugin | You structure notes with frontmatter and build dashboards by hand | vs Obsidian + Dataview |
| Nimbalyst | Visual agent workspace | You run sessions inside it and track them there | vs Nimbalyst |
| Beads | Agent-first issue tracker | Agents plan and close work in its git-synced database, per repo | vs Beads |
| Task Master | PRD-to-tasks for agents | You write a PRD; it generates tasks your agent executes | vs Task Master |
| AgentsView | Agent session analytics | It indexes session transcripts - activity and cost, not doc state | vs AgentsView |
| Plain VS Code + CLAUDE.md | The free DIY default | You read the files and keep the state in your head | vs VS Code + CLAUDE.md |
| Just asking your agent | The null alternative - a prompt, not a tool | The agent re-reads the docs and composes a summary, per ask, per project | vs just asking your agent |
Markdown viewers are deliberately absent: sitrep renders markdown at full fidelity, but a viewer answers "show me this file" - the job here is "where do my projects stand".
Measured, not asserted
sitrep is built over a real working corpus: roughly 54,900 markdown files across four workspaces, of which about 730 are actual project documents. Every extraction rule is measured against that corpus before it ships, and every UI claim is verified by driving the real app in headless Chrome and reading numbers back.
The same standard applies to what we say here. Claims stay concrete and checkable - which agents it reads, what it extracts, what it refuses to do - and the comparison pages are honest about when you should pick the other tool.
Security & privacy, and how to verify itWho builds sitrep, and against what corpus
Common questions
- Do I need to change my files?
- No. No frontmatter, no format, no import. sitrep reads what your agents already write.
- Is my data uploaded anywhere?
- No. sitrep is local-first and fully functional with no account and no network. No telemetry by default.
- Is it a markdown viewer?
- No - it renders markdown well, but the product is the state layer: status, decisions and staleness across projects.