PHPackages sellinnate/rag-engine - PHPackages - PHPackages [Skip to content](#main-content)[PHPackages](/)[Directory](/)[Categories](/categories)[Trending](/trending)[Leaderboard](/leaderboard)[Changelog](/changelog)[Analyze](/analyze)[Collections](/collections)[Log in](/login)[Sign up](/register) 1. [Directory](/) 2. / 3. [API Development](/categories/api) 4. / 5. sellinnate/rag-engine ActiveLibrary[API Development](/categories/api) sellinnate/rag-engine ===================== Enterprise Retrieval-Augmented Generation engine for Laravel: ingestion, parsing, chunking, embedding, vector store, retrieval, reranking, BYOK security and multi-tenancy. v1.2.0(today)00MITPHPPHP ^8.2CI passing Since Jun 27Pushed todayCompare [ Source](https://github.com/Sellinnate/laravel-ultimate-rag)[ Packagist](https://packagist.org/packages/sellinnate/rag-engine)[ Docs](https://github.com/Sellinnate/laravel-ultimate-rag)[ GitHub Sponsors](https://github.com/:vendor_name)[ RSS](/packages/sellinnate-rag-engine/feed)WikiDiscussions main Synced today READMEChangelog (4)Dependencies (14)Versions (5)Used By (0) [![RAG Engine for Laravel β€” semantic search & AI answers](art/banner.png)](art/banner.png) RAG Engine for Laravel ====================== [](#rag-engine-for-laravel) ![Tests](https://camo.githubusercontent.com/aec4a0df2e2d724466b3761a00bcf62755b2f120a4c12a847a48bf364ff3243c/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f74657374732d34323925323070617373696e672d627269676874677265656e)![Coverage](https://camo.githubusercontent.com/d199f8dfc0ccc91cfcf397e3f8ade5fab92a5fd35b728bb1310218d0a12ce3a9/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f636f7665726167652d25453225383925413539302532352d627269676874677265656e)![PHPStan](https://camo.githubusercontent.com/1b02b2f6c2946c9b9ad2e14b1c79c5932fdcef4ea31b4d6e79d54af7d7e7a2a7/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f5048505374616e2d6c6576656c253230382d626c7565)![PHP](https://camo.githubusercontent.com/5ed842996550cad148e16d3ad4738491b8319d2fbf7bba88eeeee0cb4bf9b736/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f5048502d382e322532422d373737626234)![Laravel](https://camo.githubusercontent.com/7df61840a0d13a570921441b8a5da7c66d91b0b27763658f25c7f861881b5cfe/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f4c61726176656c2d3131253230253743253230313225323025374325323031332d666632643230)[![License](https://camo.githubusercontent.com/f8df3091bbe1149f398a5369b2c39e896766f9f6efba3477c63e9b4aa940ef14/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d49542d677265656e)](LICENSE.md)[![Docs](https://camo.githubusercontent.com/a67a795bc38e35aa4a7b1655e5255373216e249f38f7ff5ef537b56064cbcf2b/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f646f63732d6c61726176656c2d2d7261672d2d656e67696e652e73656c6c692e696f2d323536336562)](https://laravel-rag-engine.selli.io) πŸ“– **Full documentation: [laravel-rag-engine.selli.io](https://laravel-rag-engine.selli.io)** Add **semantic search and AI answers over your own content** to any Laravel app. RAG Engine owns the whole Retrieval-Augmented Generation pipeline β€” ingesting documents, splitting them, turning them into searchable vectors, and retrieving the most relevant passages for any query. Writing a final answer with an LLM is an optional layer on top. > **Infrastructure, not a feature.** The engine owns *ingestion β†’ retrieval*; generation is optional and decoupled. Vertical packages, internal agents and search modules build on top without re-implementing ingestion, chunking, embedding or retrieval. ### What you can build [](#what-you-can-build) - πŸ”Ž **Semantic search** β€” a search box that matches by *meaning*, not keywords. - πŸ€– **AI Q&A / chatbots** β€” LLM answers grounded in *your* content, with citations. - πŸ“š **"Ask your docs / tickets / wiki"** features inside an existing app. - 🧭 **Similarity / recommendations** β€” "find records like this one". Use just the search half (no LLM, no AI bill) or add generation later β€” same code, one config switch. Table of contents ----------------- [](#table-of-contents) - [Features](#features) - [Requirements](#requirements) - [Installation](#installation) - [Quick start](#quick-start) - [Indexing Eloquent models](#indexing-eloquent-models) - [Asking questions with an LLM](#asking-questions-with-an-llm) - [Configuration](#configuration) - [Supported drivers](#supported-drivers) - [Security & multi-tenancy](#security--multi-tenancy) - [Documentation](#documentation) - [Testing & development](#testing--development) - [License](#license) Features -------- [](#features) - **Multi-format ingestion** β€” raw text, file uploads, URLs (SSRF-guarded), cloud storage and Eloquent records. Safely parses Markdown, HTML, XML, CSV, JSON, DOCX and PDF. - **Pluggable everything** β€” parsing, chunking, embedding, vector store, reranking and LLM are swappable drivers behind stable contracts. - **10 embedding providers** β€” OpenAI, Azure OpenAI, Mistral, Jina, Voyage, Cohere, Gemini, Hugging Face, Ollama, plus a deterministic `fake` driver for tests. - **Powerful retrieval** β€” metadata filters, hybrid (semantic + keyword) search with RRF, MMR diversification, reranking, relevance thresholds and small-to-big (parent-child) context expansion. - **Embeddable Eloquent models** β€” make any model searchable via one contract; recursive composition of relations, auto-sync on change, and vectorβ†’model trace-back. - **Security by design** β€” BYOK envelope encryption with a KMS abstraction (`local` + **AWS KMS**), crypto-shredding for "right to erasure", and PII redaction on by default. - **OCR for scanned PDFs** β€” pluggable OCR (Tesseract) kicks in when a PDF has no text layer. - **Quality evaluation** β€” measure recall@k, precision@k, hit-rate and MRR over a labelled dataset (`rag:evaluate`). - **Resilient providers** β€” LLM, reranker and embedder HTTP calls retry transient failures with exponential backoff. - **Multi-tenancy** β€” automatic, fail-closed per-tenant scoping of every query. - **Operations from day one** β€” immutable (WORM) audit log, cost tracking, lifecycle events, queued/batchable ingestion and Artisan commands. - **EU-resident by default** β€” content and embeddings stay in the EU unless you explicitly opt into a non-EU provider. Requirements ------------ [](#requirements) RequirementVersionPHP**8.2+**Laravel**11, 12 or 13**A databaseany Laravel-supported (SQLite is fine to start)A dedicated vector database is **not** required to begin: the default store is in-memory, and the `database` store works on plain Postgres/MySQL/SQLite. Use native `pgvector` or Qdrant at larger scale. Installation ------------ [](#installation) ``` composer require sellinnate/rag-engine php artisan vendor:publish --tag="rag-engine-config" php artisan vendor:publish --tag="rag-engine-migrations" php artisan migrate ``` The service provider and `Rag` facade auto-register via package discovery. Out of the box the package uses zero-network, deterministic drivers (`fake` embedder, in-memory store, local KMS) so your test suite runs offline. Important The `fake` embedder is for **tests only** β€” it doesn't understand meaning. For a real search feature, configure a real embedder (see [Configuration](#configuration)). Quick start ----------- [](#quick-start) ``` use Sellinnate\RagEngine\Facades\Rag; // 1. INGEST β€” register content as a Document (stored & encrypted, not yet searchable). $document = Rag::ingest( Rag::source()->text('Refunds are issued within 14 business days of an approved request.') ); // 2. PROCESS β€” run the pipeline: parse β†’ clean & redact PII β†’ chunk β†’ embed β†’ store. Rag::process($document); // or: ProcessDocumentJob::dispatch(...) on a queue // 3. SEARCH β€” find the most relevant chunks by meaning. $hits = Rag::search('how long until I get my money back?')->topK(3)->get(); $hits[0]->content; // "Refunds are issued within 14 business days..." $hits[0]->score; // relevance score $hits[0]->metadata['source_ref']; // provenance: where it came from // 4. (OPTIONAL) ASK β€” let an LLM write a cited answer from the retrieved chunks. $answer = Rag::ask('how long do refunds take?')->using('openai')->generate(); $answer->answer; // "Refunds take 14 business days. [1]" $answer->citations; // [['index' => 1, 'document_id' => '…', 'chunk_id' => '…']] ``` Refine retrieval fluently: ``` $hits = Rag::search('envelope encryption') ->topK(5) ->threshold(0.4) // drop weak matches ->where('tag', 'docs') // metadata filter ->hybrid() // semantic + keyword (RRF) ->rerank() // precision pass ->expandParents() // small-to-big context ->get(); ``` Indexing Eloquent models ------------------------ [](#indexing-eloquent-models) If the content you want to search already lives in your database, make the model embeddable β€” it then stays in sync automatically as rows change, and every vector traces back to its model. ``` use Sellinnate\RagEngine\Concerns\HasEmbeddings; use Sellinnate\RagEngine\Contracts\Embeddable; use Sellinnate\RagEngine\Eloquent\EmbeddableDefinition; class Article extends Model implements Embeddable { use HasEmbeddings; // auto-indexes on save, removes on delete public function toEmbeddable(): EmbeddableDefinition { return EmbeddableDefinition::make() ->add('Title', $this->title) ->add('Body', $this->body) ->include($this->author, 'author') // compose a related model ->includeMany($this->comments, 'comments'); // recursively } } // Trace a search hit back to its model: $article = Rag::models()->resolve($hits[0]); // App\Models\Article instance, or null ``` Model **file fields** (a PDF/DOCX upload) can be embedded too β€” `addFile()` parses the file to text and folds it into the model's embedding; non-embeddable binaries (zip/exe…) are skipped or rejected per policy. See **[docs/concepts/eloquent-models.md](docs/concepts/eloquent-models.md)**. Asking questions with an LLM ---------------------------- [](#asking-questions-with-an-llm) Search returns the relevant chunks; an **LLM** turns them into a written, cited answer via `Rag::ask()`. This layer is optional and decoupled β€” with the default `null` driver, `ask()` returns the sources with an empty answer, so search-only apps carry no LLM dependency. The package ships two LLM drivers: **`anthropic`** (Claude) and **`openai`**(OpenAI and any OpenAI-compatible API β€” Mistral, Ollama, Groq, OpenRouter…). ``` # .env β€” use Anthropic Claude to answer questions RAG_LLM=anthropic RAG_ANTHROPIC_API_KEY=sk-ant-... RAG_ANTHROPIC_MODEL=claude-sonnet-4-6 # or claude-opus-4-8 / claude-haiku-4-5-... ``` ``` $result = Rag::ask('What is our refund policy?') ->topK(5) ->using('anthropic') // or omit to use the default RAG_LLM ->generate(); $result->answer; // "Refunds are issued within 14 business days. [1]" $result->citations; // [['index' => 1, 'document_id' => '…', 'chunk_id' => '…']] $result->sources; // the SearchHits the answer was built from ``` Note **Anthropic has no embedding API**, so `anthropic` is a *generation-only*driver. Keep a real `RAG_EMBEDDER` (Mistral, OpenAI, Ollama…) for the search side. A common combo is **Mistral/Ollama embeddings + Claude answers**. Retrieved content is treated as untrusted: the default prompt fences it and tells the model not to follow instructions inside it (prompt-injection hardening). Full guide: **[docs/concepts/generation.md](docs/concepts/generation.md)**. Configuration ------------- [](#configuration) Configuration lives in `config/rag-engine.php` and works like Laravel's `config/database.php`: you define named *connections* per subsystem and pick a `default`. Switching provider = changing one name in `.env`. ``` # .env β€” switch to a real embedder (Ollama is free & local) RAG_EMBEDDER=ollama RAG_OLLAMA_BASE_URL=http://localhost:11434 # ...or a hosted provider: RAG_EMBEDDER=openai RAG_OPENAI_API_KEY=sk-... ``` Note **API keys go in `.env`, never in the committed config.** A copy-ready list of every variable ships as [`.env.example`](.env.example). See [docs/getting-started/configuration.md](docs/getting-started/configuration.md). Supported drivers ----------------- [](#supported-drivers) **Embedders** (`RAG_EMBEDDER`) DriverProviderResidency`openai`OpenAIglobal`azure-openai`Azure OpenAIEU (EU region)`mistral`MistralEU`jina`Jina AIEU`voyage`Voyage AIglobal`cohere`Cohereglobal`gemini`Google Geminiglobal`huggingface`Hugging Face / self-hosted TEIglobal / self-host`ollama`Ollama (BGE/E5/Nomic)self-hosted`fake`deterministic (tests)local**Vector stores** (`RAG_VECTOR_STORE`): `memory` (tests/dev) Β· `database`(portable SQL: Postgres/MySQL/SQLite, brute-force) Β· `pgvector` (**native**Postgres ANN: `vector` column + HNSW + ``) Β· `qdrant` (EU self-hostable, ANN at scale). Full setup, including **where to configure the Postgres connection**, is in the [Vector stores guide](https://laravel-rag-engine.selli.io/concepts/vector-stores). **LLMs** (`RAG_LLM`, for `ask()`): `anthropic` (Claude) Β· `openai` (OpenAI and any OpenAI-compatible API: Mistral, Ollama, Groq, OpenRouter…) Β· `null`/`fake`. Anthropic is generation-only (no embeddings). Answers can be **streamed** with `Rag::ask(...)->stream()`. **Rerankers** (`RAG_RERANKER`, optional cross-encoder pass): `cohere` Β· `jina`(EU) Β· `null`/`fake`. **KMS** (`RAG_KMS`, BYOK key management): `local` (dev) Β· `aws` (AWS KMS, production). **OCR** (`RAG_OCR`, scanned-PDF fallback): `null` Β· `tesseract`. **Parsers**: plain text Β· Markdown Β· HTML Β· XML Β· CSV/TSV Β· JSON Β· DOCX Β· PDF (+ OCR for scans). **Chunkers**: `recursive` (default) Β· `sentence` Β· `markdown` Β· `fixed`(char- or token-based), with optional parent-child and contextual headers. All drivers share one contract β€” switching backends needs no code changes, and you can register your own (see [docs/guides/custom-drivers.md](docs/guides/custom-drivers.md)). Security & multi-tenancy ---------------------------- [](#security--multi-tenancy) - **BYOK envelope encryption** β€” content is encrypted at rest with per-item DEKs wrapped by a tenant KEK in a KMS; the plaintext key never persists. - **Crypto-shredding** β€” honour "right to erasure" by destroying the key, making data unrecoverable everywhere (including DB backups) at once. - **PII redaction** β€” emails, cards (Luhn), IBANs (mod-97), Italian fiscal codes and phone numbers are redacted before indexing, by default. - **Fail-closed multi-tenancy** β€” every query is automatically scoped to the current tenant; scope can never be widened from a query (a tested invariant). - **Tamper-evident audit log** β€” append-only with database-level WORM triggers. ``` use Sellinnate\RagEngine\Facades\Rag; // Run work scoped to a tenant (previous tenant restored afterwards): Rag::forTenant('tenant-7', fn () => Rag::search('q')->get()); // Right to erasure β€” crypto-shred a tenant: Rag::kms()->destroyKey('tenant-7'); ``` See [docs/concepts/security.md](docs/concepts/security.md) and [docs/concepts/multi-tenancy.md](docs/concepts/multi-tenancy.md). Documentation ------------- [](#documentation) The full documentation is hosted at **[laravel-rag-engine.selli.io](https://laravel-rag-engine.selli.io)**. The sources live in [`docs/`](docs/) and are built into a static site with [docmd](https://docmd.io): ``` npm install npm run docs:dev # local preview npm run docs:build # static site into ./site ``` **Start here:** - 🧠 [What is RAG?](https://laravel-rag-engine.selli.io/getting-started/what-is-rag) β€” concepts + glossary, from zero. - πŸš€ [Quickstart](https://laravel-rag-engine.selli.io/getting-started/quickstart) β€” a complete worked example. - πŸ—οΈ [Architecture](https://laravel-rag-engine.selli.io/concepts/architecture) β€” how the pieces fit together. - πŸ“₯ [Ingesting content](https://laravel-rag-engine.selli.io/guides/ingestion) Β· πŸ”Ž [Retrieval & search](https://laravel-rag-engine.selli.io/concepts/retrieval) Β· πŸ’¬ [Generation](https://laravel-rag-engine.selli.io/concepts/generation) - πŸ—„οΈ [Vector stores & configuration](https://laravel-rag-engine.selli.io/concepts/vector-stores) β€” incl. pgvector Postgres setup. - πŸ“Š [Evaluating quality](https://laravel-rag-engine.selli.io/guides/evaluation) β€” recall@k, MRR, `rag:evaluate`. - 🧩 [Contracts reference](https://laravel-rag-engine.selli.io/reference/contracts) Β· πŸ› οΈ [Custom drivers](https://laravel-rag-engine.selli.io/guides/custom-drivers) Testing & development ------------------------- [](#testing--development) ``` composer test # run the Pest suite (429 tests) composer analyse # PHPStan, level 8 composer format # Laravel Pint (code style) # Coverage (needs a coverage driver, e.g. Xdebug/PCOV): XDEBUG_MODE=coverage vendor/bin/pest --coverage --min=90 ``` Quality gates kept green on every change: **429 tests**, **PHPStan level 8**, **Pint** clean, **β‰₯90% coverage**. License ------- [](#license) MIT β€” see [LICENSE.md](LICENSE.md). ### Health Score 41 β€” FairBetter than 87% of packages Maintenance100 Actively maintained with recent releases Popularity0 Limited adoption so far Community6 Small or concentrated contributor base Maturity49 Maturing project, gaining track record Bus Factor1 Top contributor holds 100% of commits β€” single point of failure How is this calculated?**Maintenance (25%)** β€” Last commit recency, latest release date, and issue-to-star ratio. Uses a 2-year decay window. **Popularity (30%)** β€” Total and monthly downloads, GitHub stars, and forks. Logarithmic scaling prevents top-heavy scores. **Community (15%)** β€” Contributors, dependents, forks, watchers, and maintainers. Measures real ecosystem engagement. **Maturity (30%)** β€” Project age, version count, PHP version support, and release stability. ### Release Activity Cadence Every ~0 days Total 4 Last Release 0d ago ### Community Maintainers ![](https://www.gravatar.com/avatar/8df9dced104a057028a8e34fb75f3206562b0d0d700fb2dcaea834e490a6cdc8?d=identicon)[sellinnate](/maintainers/sellinnate) --- Top Contributors [![FilippoCalabrese](https://avatars.githubusercontent.com/u/15164324?v=4)](https://github.com/FilippoCalabrese "FilippoCalabrese (52 commits)") --- Tags laravelembeddingsqdrantragpgvectorretrieval-augmented-generationvector-searchSellinnate ### Code Quality TestsPest Static AnalysisPHPStan Code StyleLaravel Pint ### Embed Badge ![Health badge](/badges/sellinnate-rag-engine/health.svg) ``` [![Health](https://phpackages.com/badges/sellinnate-rag-engine/health.svg)](https://phpackages.com/packages/sellinnate-rag-engine) ``` ### Alternatives [dedoc/scramble Automatic generation of API documentation for Laravel applications. 2.1k9.9M90](/packages/dedoc-scramble)[spatie/laravel-pdf Create PDFs in Laravel apps 1.0k4.3M42](/packages/spatie-laravel-pdf)[defstudio/telegraph A laravel facade to interact with Telegram Bots 815320.5k3](/packages/defstudio-telegraph)[rawilk/profile-filament-plugin Profile & MFA starter kit for filament. 3913.7k](/packages/rawilk-profile-filament-plugin)[simplestats-io/laravel-client Analytics for Laravel. Track visitors, registrations, and payments. Discover which channels actually drive revenue, not just traffic. Server-side, GDPR compliant, ad-blocker proof. 5019.3k](/packages/simplestats-io-laravel-client)[spatie/laravel-github-webhooks Handle GitHub webhooks in a Laravel application 93157.3k5](/packages/spatie-laravel-github-webhooks) PHPackages Β© 2026 [Directory](/)[Categories](/categories)[Trending](/trending)[Changelog](/changelog)[Analyze](/analyze)