senecaAIby Vindica
Menu
ProductDevelopersPilotTrustCompanyInvestorsRequest a pilot한국어

Build with Seneca

Give your product a record that travels.

Your product keeps its curriculum, teaching logic and interface. Seneca carries the observable evidence underneath so the learner does not begin from zero in the next system.

Map one item eventChoose an integration

See it work

Watch a learner’s history reach the next product.

A fictional student moves between an assessment app at school, a tutoring app on their phone and a reading app on a tablet. Seneca carries what one product observed to the next product the student allows. Each product still decides what to teach.

  1. 1One product learns something. An assessment app records fractions, word problems, what helped and what looked rushed.
  2. 2Seneca keeps the evidence. Each conclusion shows its source, date, amount of evidence and whether it still needs checking.
  3. 3The student decides. A tutoring app asks to read the maths context. A private reading journal stays locked.
  4. 4The next product starts informed. Switch Seneca off and on to compare the tutoring app’s first minute, then follow its result back to the assessment app.
Open the interactive demo

Fictional student and apps. About two minutes to click through.

Starter kit

Stop re-testing every new learner.

Follow one made-up student from a test app to a tutoring app in four plain steps. Engineers get a technical version with the real files, a browser tool for their own made-up data, and a download.

Open the starter kit

Choose the connection

Start self-service, or become a product limb.

These are different integration shapes. An assistant connects one consenting learner. A limb connects a learning product’s item and session events to the record.

Live · self-service read

Connect an AI assistant

Use the MCP endpoint to read a learner-authorized summary before a session. Approved write-capable clients can record the completed session afterward.

  • OAuth discovery and dynamic client registration
  • Read access is self-service; write access is reviewed
  • Per-learner consent and immediate revocation
https://vindicaseneca.com/mcp

Reviewed · pilot

Connect your product as a limb

Map your existing right/wrong, timing, assistance and item-context data into Seneca. The SDK handles the shared record; your product remains yours.

  • Server-to-server writes with source attribution
  • Mode-specific schema and observable item passport
  • Integration baseline, audit trail and scoped reads
Request a limb pilot

The first design task

Map one answered item into an observable passport.

Start with facts your product already has. Change the fields below to see the portable evidence packet. This does not assign a universal ability score; it preserves what happened and the conditions under which it happened.

Your taxonomy stays intact. The limb mapping links it to a Seneca mode schema.
Record preview

Conceptual item passport. During a pilot, we turn this mapping into the validated production envelope for your limb and document every derived field.

Production path

Four steps from local event to portable evidence.

1

Map

Pair your local item types and observable fields with the right Seneca mode schema.

2

Link

Connect the learner’s product identity to the record with explicit authorization.

3

Write

Submit item and session events through the SDK. Seneca records source and provenance.

4

Read

Return a minimized, scoped starting point to another product the learner authorizes.

What can travel

  • Outcome, timing and attempt history
  • Assistance state and testing conditions
  • Observable item and task features
  • Source, provenance, scope and uncertainty

What stays local

  • Your curriculum and teaching sequence
  • Your interface and recommendation logic
  • Your proprietary questions and content
  • Anything the learner did not authorize

Live now

Connect an assistant over MCP.

The assistant connection is the fastest way to inspect Seneca’s permission model. Add the endpoint; discovery, registration and learner consent follow over OAuth. Self-service clients receive narrowed read access; writing requires an approved client.

whoamiConfirm the connected learner and granted scopes.
get_learning_summaryRead the deliberately narrowed longitudinal summary.
get_ability_mapRead demonstrated abilities and explicit unmeasured areas in the supported domain.
record_learning_sessionWrite the completed interaction back with source attribution.
MCP endpoint
https://vindicaseneca.com/mcp          # asks: identity.basic bridge.read.shared
https://vindicaseneca.com/mcp/record   # asks: identity.basic record.read.summary bridge.read.shared

# Same server at two addresses; the learner picks the consent.
# record.write.observation is asked by neither — approved clients only.

# Learner control
authorization: per learner
revocation: immediate
raw question history: not exposed

Already writing

Six first-party limbs exercise the record today.

They span standardized tests, speaking, writing and job-search assessment. The next proof is an outside product completing the same write-and-read path.

FORGE · Digital SATMaroTOS마로 인적성MaroPrepKAAT · AI 활용 평가VindicaOS

First external limb

Bring one real item type.

We will map the observable fields, define what must remain local, establish an onboarding baseline and scope a contained production pilot.

Request a pilot