cognesy / instructor-polyglot

v2.10.1 2026-09-13 07:31 UTC

Requires

• php: ^8.3
• cognesy/instructor-config: ^2.10
• cognesy/instructor-events: ^2.10
• cognesy/instructor-http-client: ^2.10
• cognesy/instructor-messages: ^2.10
• cognesy/instructor-metrics: ^2.10
• cognesy/instructor-setup: ^2.10
• cognesy/instructor-telemetry: ^2.10
• cognesy/instructor-utils: ^2.10
• cognesy/logging: ^2.3

Requires (Dev)

• icanhazstring/composer-unused: ^0.9.0
• jetbrains/phpstorm-attributes: ^1.2
• maglnet/composer-require-checker: ^4.16
• mockery/mockery: ^1.6
• pestphp/pest: ^4.0
• phpstan/phpstan: ^1.11
• roave/security-advisories: dev-latest
• vimeo/psalm: ^6.0

Suggests

• cognesy/instructor-addons: Provided extra optional capabilities like evals, tool use, etc.

MIT

Published on Packagist.org on 2026-09-13 07:36 UTC

README

Unified LLM connectivity layer for InstructorPHP.

It provides two facades:

• Inference for chat/completion responses
• Embeddings for vector generation

Example

<?php

use Cognesy\Polyglot\Inference\Inference;
use Cognesy\Polyglot\Inference\Reasoning\ReasoningEffort;
use Cognesy\Polyglot\Inference\Reasoning\ReasoningSelection;

$message = Inference::using('openai')
    ->withModel('gpt-5.6')
    ->withMessages('Write one short sentence about PHP.')
    ->withReasoning(ReasoningSelection::effort(ReasoningEffort::Medium))
    ->get();

echo $message->content()->toString();

Documentation

• packages/polyglot/docs/quickstart.md
• packages/polyglot/docs/essentials/inference-class.md
• packages/polyglot/docs/essentials/reasoning.md
• packages/polyglot/docs/embeddings/overview.md
• packages/polyglot/docs/_meta.yaml