Built for Filament 5+

Laravel RAG Chatbot PluginA chatbot that cites its sources.

Point it at your docs, URLs and PDFs. RAG Chatbot turns them into grounded answers with citations, and you run every bot from your Filament panel.

PHP 8.4+Laravel 12+
Live widget · example datafig. 01

Docs Assistant

Answers from acme.dev/docs

● online
How do I limit the widget to our own domain?
Add your domain to the bot’s allowlist under Widget → Security. Requests from other origins are rejected and every embed is signed with a token1. You can also cap requests per minute2.

SOURCES

1widget/security.md0.91
2operations/rate-limits.md0.84
Ask the docs…

RETRIEVAL

top-k6
min. similarity0.72

Bring your own model

  • Gemini
  • OpenAI
  • Anthropic
  • xAI
  • pgvector
  • Chroma

// 01 — How it works

From a folder of docs to answers you can trust.

Four steps, all visible in your panel. You can see what was indexed, what was retrieved and why the bot answered the way it did.

  1. STEP 1

    Ingest

    Text, files and URLs. Markdown, HTML and PDFs go through the queue, with retries.

    handbook.pdfindexed
    acme.dev/docs/*indexed
    pricing.mdqueued
  2. STEP 2

    Embed & store

    Chunks become vectors in PostgreSQL with pgvector, or in Chroma if you prefer.

  3. STEP 3

    Retrieve

    Tune top-k, similarity threshold and context budget per bot, without a deploy.

    0.91
    0.84
    0.76
    0.61

    — threshold 0.72 —

  4. STEP 4

    Answer with citations

    Every answer links back to the passage it came from, so your users can check.

    Refunds are processed within 14 days1.

// 02 — Control room

Everything lives in your Filament panel.

One panel, many bots

Own prompt, model and sources per bot

BOTMODELAUDIENCESTATUS
Docs AssistantGeminipublic● live
Customer SupportOpenAImembers● live
Internal CopilotAnthropicadmins● indexing

Health check, built in

$ php artisan filament-rag:doctor

✓ vector store reachable

✓ queue worker running

✓ provider key valid

All checks passed.

Embed with one line

A script tag for any site, an NPM loader for SPAs.

<script src="…/widget.js"
  data-bot="docs-assistant"></script>

Locked to your domains

Signed tokens, an allowlist and rate limits per bot.

acme.devapp.acme.dev+ add domain

Review every conversation

See which answers had sources, and where your docs fall short.

“Can I self-host?”2 sources
“SSO with Azure AD?”no source

// 03 — Real screens

Straight from the live demo.

Open the demo
RAG Chatbot widget settings in Filament with a live preview of the chat widget per area
fig. 02Widget settings with a live preview per area: public, member or admin, each with its own look.
Table of RAG bots in the Filament panel with provider, vector backend, model and sources
fig. 03Every bot in one Filament table: provider, vector backend, model and sources.
Edit RAG bot page with embed snippet, setup check, analytics and diagnostics actions
fig. 04Bot settings with the embed snippet, setup check, analytics and diagnostics one click away.

// 04 — Use cases

Where teams use it.

  1. 01

    AI documentation chatbot for Laravel products

    Turn product docs, help articles, and Markdown content into grounded answers with retrieval context and citations.

  2. 02

    Customer support chatbot with a Filament control panel

    Let support or product teams manage sources, review conversations, and adjust bot settings without direct database work.

  3. 03

    pgvector chatbot for Laravel and Filament

    Use PostgreSQL with pgvector or ChromaDB to run retrieval-backed chat while keeping admin operations in one panel.

// 05 — Right fit?

Answers only, or answers and actions?

RAG Chatbot

You need answers from your docs.

  • Teams that need grounded support or documentation chatbots
  • Products already standardized on Laravel, Filament, and pgvector/Chroma
  • Founders who want a proven admin UX instead of wiring LangChain glue code

Agentic Chatbot

Your bot also needs to act.

  • Visual workflows with branches and human handoff
  • API connectors and safe database lookups
  • Telegram, Slack and full execution traces
See Agentic Chatbot

// 06 — License

Install it today.

Licenses are sold through the official checkout. You get private Composer access and updates according to the published license terms.

INCLUDED

  • Private Composer repository
  • Updates per the license terms
  • Full documentation
  • A public demo to test first

REQUIRES

  • PHP 8.4+
  • Laravel 12+
  • Filament 5+
  • PostgreSQL + pgvector (recommended) or Chroma
  • LLM provider API key (Gemini, OpenAI, Anthropic, xAI, etc.)

// 07 — FAQ

Questions before you buy.

Something missing? I built the plugin, so ask me directly.

Send a question
How is this different from generic ChatGPT integrations?

Filament RAG focuses on your data: ingestion pipelines, retrieval tuning, citations, and Filament resources for day-two operations—not a single hard-coded prompt.

Can I run multiple bots?

Yes. Each bot can have its own sources, models, prompts, and widget branding, which is ideal for multi-product suites or customer-specific knowledge spaces.

Is there a health check command?

Run `php artisan filament-rag:doctor` after configuring your vector backend and queue to validate connectivity and core requirements.

Where can I read the full documentation?

Public docs live at filament-rag.heinerdevelops.tech/docs with quickstart, operations, and security guides.

// 08 — Field notes

Read before you build.

All articles

// More from the workshop

All plugins

Your docs could be answering by tonight.