Home

Causet turns fragile workflows into replayable timelines.

Causet sits beside your existing app. It does not replace your frontend, database, queue, vector DB, or services. Your app submits intents. Causet records events. Queries and projections expose current state. The timeline can be inspected, forked, replayed, or repaired.

Adopt it one endpoint, webhook, background job, or agent action at a time.


Start here

The shortest path to one successful first workflow:

  1. Understand the mental model — state, intent, event, query, projection, timeline, replay, fork, repair
  2. Run the browser tutorial — learn the model in ~5 minutes, no signup
  3. Retrofit one existing endpoint — keep the API contract; hand the workflow to Causet
  4. Inspect the timeline — see what happened and why
  5. Simulate a failure — make a broken step visible
  6. Replay or repair — fix downstream state without guessing from logs

Existing app? Start at step 3 after a quick pass through the mental model. Retrofit is the primary adoption path.


Docs map

SectionStart with
What is Causet?What is Causet? · What runs today? · When not to use Causet
The 5-minute browser tutorial5-Minute Quickstart
Retrofit an existing endpointRetrofit an Existing App
Webhooks and async workflowsUse Webhooks to Update Existing Flows
AI decisions and memoryAI in Causet — additive, not the only use case
Timeline, replay, and repairTimeline · Replay
CLI and compilerInstall Causet
Runtime and deploymentRun Causet with a Local App · Deployments
ReferenceDSL · Examples

What runs today?

CapabilityStatus
Browser tutorialAvailable now
CLI validate / compileAvailable now
Timeline inspectorAvailable now (Cloud early access)
Replay / fork / repairAvailable now (Cloud early access)
Managed CloudEarly access
SDK packagesPreview
Local runtime / DockerComing soon
Outbound HTTP webhook deliveryAvailable now

Full table and notes: What runs today?.


Why teams add Causet

API routes, webhook handlers, and background jobs often hide multi-step workflows: validate, write DB, call a service, send email, update status — then fail halfway with no clear timeline.

Causet gives those workflows:

  • a durable timeline of meaningful state changes
  • inspect, fork, replay, and repair tools
  • a way to keep the existing frontend reading existing tables via projections or webhook/event bridges

AI agents make the same problem sharper — decisions, memory, tool calls, and side effects also need to be inspectable. AI is not the only use case. It is the sharpest version of the problem.


One concrete win

Before: POST /api/access-requests validates, writes a row, calls approval, sends email, updates status — and maybe fails halfway.

After: the endpoint validates and submits request_access. Causet records intents and events (access_requested, approval_required, access_approved, notification_sent). Your app updates its own DB from Causet events or projections. Failures show up in the timeline and can be replayed.

Walk through it: Retrofit an Existing App.


You still own

  • Frontend and API contracts
  • Primary database (if you choose)
  • Queues, workers, and third-party clients
  • Vector DB and other services

Causet is the deterministic state layer for workflows where correctness, auditability, replay, and repair matter — not a replacement for your stack.


Browse by need