P/CPromptCapsule Sign in

Intent Engineering infrastructure

Engineer Intent.
Preserve Intelligence.

PromptCapsule transforms prompts from disposable text into persistent, governed communication matter—capturing human intent, enabling reproducible AI reasoning, and evolving organizational intelligence over time.

Enter the workspaceExamine a capsule ↓
Version controlledModel independentMeasurable fidelity
PC-1842 / v3.2
INTENT ARTIFACT

Architecture decision synthesis

APPROVED
Intent fidelity92
Purpose
Preserve the decision logic—not only its conclusion.
Constraints
Evidence-linked · reversible · tenant-scoped
Evaluation
Traceability 40% · completeness 35% · clarity 25%
v3.214 executions3 linked capsules

Prompts are not ephemeral instructions.

They are the source code of human–AI collaboration.

Source code has repositories, history, tests, owners, and controls. Intent deserves the same.

01 / Intent Engineering

Make intended reasoning reconstructable.

Intent Engineering is the discipline of designing, refining, governing, and measuring communication artifacts so another intelligence can reconstruct not only the request, but the intended reasoning.

01

Intent initiates

A purpose begins the creation of communication matter.

02

Prompts encode

Purpose, context, constraints, assumptions, and evaluation become persistent.

03

Reasoning traverses

An intelligence reconstructs associations and traverses a Communication Association Graph.

04

Evaluation returns

Outputs produce evidence that clarifies, challenges, and refines the original intent.

CFT CANDIDATE LAW

Intent fidelity sets the ceiling.

The precision with which intent is encoded determines the maximum achievable fidelity of the reasoning process—and how objectively its outcome can be evaluated.

Read the theoretical foundation ↗

02 / Anatomy

A governed package for intent.

A PromptCapsule carries everything required to understand, execute, reproduce, evaluate, and improve an intent artifact.

PROMPTCAPSULEIntentPersistent communication matter
  1. DefineTitle · purpose · original intent · expected outcome
  2. FramePrompt · context · constraints · attachments
  3. ExecutePersonas · models · runtime · cost · outputs
  4. EvaluateRubric · feedback · satisfaction · comparisons
  5. EvolveVersions · derivations · relations · tags · projects

Git manages source code.PromptCapsule manages intent.Preserve not only what was built, but why.

03 / Lifecycle

Evidence makes intent better.

  1. 01IntentPurpose to preserve
  2. 02EncodePrompt + conditions
  3. 03TraverseAI reasoning
  4. 04GenerateNew matter
  5. 05EvaluateExplicit rubric
  6. 06RefineVersioned intent
v2.1 / BEFORE

“Create a rollout plan for the new system.”

Intent fidelity 54
v3.0 / REFINED

“Design a reversible 30-day rollout, preserving service continuity and defining adoption, error-rate, and rollback thresholds.”

Intent fidelity 91

04 / Platform

Infrastructure for the full life of intent.

01

Capsule editor

Structure intent, context, constraints, rubrics, and expected outcomes.

02

Version & compare

Trace intent evolution and compare prompts, outputs, cost, and satisfaction.

03

Execution history

Reproduce runs across models and personas with full operational context.

04

Approval workflows

Route high-impact intent through accountable review and release controls.

05

Semantic graph search

Find capsules by purpose, associations, evidence, projects, and outcomes—not keywords alone.

06

Analytics

Measure fidelity, satisfaction, cost, runtime, reuse, and organizational improvement.

07

Collaboration & RBAC

Tenant-aware teams, permissions, sharing, audit trails, and controlled templates.

08

APIs & portability

Create, execute, export, import, and integrate capsules into existing workflows.

05 / Architecture

The intent layer for humans, agents, and enterprise knowledge.

PromptCapsule operationalizes CFT on NodeRunner infrastructure while preserving clear product boundaries.

PROMPTCAPSULEIntent artifacts · governance · fidelity · evolution
NODEVERTEXCommunication fabricWorkspaces · vertices · collaboration · orchestration · audit
1DBPersistent relationshipsVersions · outputs · evaluations · associations · knowledge graphs
NODERUNNER PLATFORMIdentity · tenants · permissions · platform APIs · enterprise integration

Every PromptCapsule is represented as a governed communication artifact within NodeVertex. Its durable state and graph relationships are persisted through 1DB’s memory fabric.

06 / Enterprise

Organizational intent becomes an asset.

Requirements, policies, playbooks, specifications, research hypotheses, and AI prompts share an architectural category: communication artifacts that encode expected meaning, action, limits, and accountability.

Governance without friction

Ownership, review, approval, retention, export, and audit controls travel with the artifact.

Institutional continuity

Teams preserve the rationale and evidence behind decisions even as people, tools, and models change.

Model independence

Measure the artifact separately from any model, then compare how different intelligences reconstruct it.

07 / Developers & API

Intent belongs in the delivery pipeline.

Use capsules as versioned inputs to agents, workflows, evaluations, and release gates.

POST /api/v1/capsules
{
  "title": "Architecture decision synthesis",
  "purpose": "Preserve decision logic and evidence",
  "intent": { "expectedOutcome": "Reconstructable ADR" },
  "constraints": ["evidence-linked", "reversible"],
  "evaluation": { "rubric": "adr-fidelity-v2" },
  "relationships": ["project:atlas", "capsule:PC-1731"]
}
API reference ↗Capsule schema ↗Import & export ↗

08 / Pricing

Start with an artifact. Scale to an institution.

INDIVIDUAL

Free

$0

Build and version personal intent artifacts.

  • 50 capsules
  • Version history
  • Export
Start engineering
TEAM

Collaborative

$24/user/mo

Govern shared intent across a working team.

  • Unlimited capsules
  • Approvals & comparisons
  • Analytics & APIs
Talk to us
ENTERPRISE

Institutional

Custom

Operate intent as governed infrastructure.

  • Enterprise SSO & RBAC
  • Private integrations
  • Retention & audit controls
Design a deployment

09 / Ecosystem

One intent layer. Many forms of intelligence.

PromptCapsuleIntent layer
CFTNodeVertex1DBVertex WorksCRMAI PersonasKnowledge Graphs

10 / Field notes

A category needs a working vocabulary.

Research notes on intent fidelity, artifact governance, reasoning reproducibility, and communication architecture.

Begin with Intent Engineering ↗

11 / Contact

Build an intent system.

For platform architecture, enterprise deployment, research collaboration, or early access.

[email protected]