Intentum.AI.OpenAI 0.0.1

Intentum - Intent-Driven Development for the AI Era

Intentum replaces scenario-based BDD with behavior space inference. It focuses on what the user was trying to do, not just what they did.

English | Türkçe

Why Intentum?

• Non-deterministic flows are now common.
• AI-driven systems adapt and drift.
• Assertions are too rigid; intent is more realistic.

When not to use

• Fully deterministic, low-variance systems with stable requirements.
• Small scripts or one-off utilities where behavior drift is irrelevant.

BDD vs Intentum

• BDD: Scenario-driven, deterministic, pass/fail
• Intentum: Behavior-driven, probabilistic, policy decisions

Documentation

• GitHub Pages (EN/TR): https://keremvaris.github.io/Intentum/
• English docs: docs/en/index.md
• Turkish docs: docs/tr/index.md
  • Enable in GitHub: Settings → Pages → Source: GitHub Actions
• API reference (auto): https://keremvaris.github.io/Intentum/api/

Core Concepts Behavior Space Observe behavior instead of writing scenarios.

Intent Inference Infer intent from the behavior space and compute confidence.

Policy Decisions Replace pass/fail with policy-driven decisions.

Quick Example

using Intentum.AI.Mock;
using Intentum.AI.Models;
using Intentum.AI.Similarity;
using Intentum.Core;
using Intentum.Core.Behavior;
using Intentum.Runtime.Policy;

var space = new BehaviorSpace()
    .Observe("user", "login")
    .Observe("user", "retry")
    .Observe("user", "submit");

var intentModel = new LlmIntentModel(
    new MockEmbeddingProvider(),
    new SimpleAverageSimilarityEngine());

var intent = intentModel.Infer(space);

var policy = new IntentPolicy()
    .AddRule(new PolicyRule(
        "HighConfidenceAllow",
        i => i.Confidence.Level is "High" or "Certain",
        PolicyDecision.Allow));

var decision = intent.Decide(policy);

Run the sample

dotnet run --project samples/Intentum.Sample

Showcase output (trimmed)

=== INTENTUM SCENARIO: PaymentHappyPath ===
Events            : 2
Intent Confidence : High
Decision          : Allow
Behavior Vector:
 - user:login = 1
 - user:submit = 1

=== INTENTUM SCENARIO: PaymentWithRetries ===
Events            : 4
Intent Confidence : Medium
Decision          : Observe
Behavior Vector:
 - user:login = 1
 - user:retry = 2
 - user:submit = 1

=== INTENTUM SCENARIO: SuspiciousRetries ===
Events            : 4
Intent Confidence : Medium
Decision          : Block
Behavior Vector:
 - user:login = 1
 - user:retry = 3

Packages

• Intentum.Core
• Intentum.Runtime
• Intentum.AI
• Intentum.AI.OpenAI
• Intentum.AI.Gemini
• Intentum.AI.Claude
• Intentum.AI.Mistral
• Intentum.AI.AzureOpenAI

Configuration (env vars)

• OPENAI_API_KEY, OPENAI_EMBEDDING_MODEL, OPENAI_BASE_URL
• GEMINI_API_KEY, GEMINI_EMBEDDING_MODEL, GEMINI_BASE_URL
• MISTRAL_API_KEY, MISTRAL_EMBEDDING_MODEL, MISTRAL_BASE_URL
• AZURE_OPENAI_ENDPOINT, AZURE_OPENAI_API_KEY, AZURE_OPENAI_EMBEDDING_DEPLOYMENT, AZURE_OPENAI_API_VERSION

Security

• Never commit API keys. Use environment variables or secret managers.
• Avoid logging raw requests/responses from providers in production.

Note AI adapters are deterministic stubs in v1.0. Real HTTP calls are planned for v1.1.

Releasing and versioning

• Versioning is automatic from git tags via MinVer (src/Directory.Build.props). No manual version bump in code: tag v1.0.1 and build → package version is 1.0.1. Tag prefix is v; minimum major.minor is 1.0.
• Pushing a tag like v1.0.1 triggers GitHub Release and NuGet publish (see .github/workflows/).
• Release notes are generated from conventional commits via git-cliff (cliff.toml). Use prefixes: feat:, fix:, docs:, chore:, ci:, etc. If none are found, CHANGELOG.md is used as fallback.
• On every push to master, the changelog workflow updates CHANGELOG.md from conventional commits and commits it (with [skip ci]).

License MIT