haiku.rag

Name: haiku.rag
Author: ggozad

5.7kinstalls/mo · pypi547stars35forks

How to install haiku.rag?

ggozad/haiku.rag

$git clone https://github.com/ggozad/haiku.rag

Installs into the current project.

Summary

Agentic RAG built on LanceDB, Pydantic AI, and Docling.

Files · 1

View on GitHub

README.md

1	# Haiku RAG
2
3	[![Tests](https://github.com/ggozad/haiku.rag/actions/workflows/test.yml/badge.svg)](https://github.com/ggozad/haiku.rag/actions/workflows/test.yml)
4	[![codecov](https://codecov.io/gh/ggozad/haiku.rag/graph/badge.svg)](https://codecov.io/gh/ggozad/haiku.rag)
5
6	Agentic RAG built on [LanceDB](https://lancedb.com/), [Pydantic AI](https://ai.pydantic.dev/), and [Docling](https://docling-project.github.io/docling/).
7
8	> New: vision and multimodal search. Picture-aware ingestion captures embedded figure bytes; vision-capable QA models receive them alongside text. Multimodal embedders put picture vectors in the same space as text, enabling text-as-query → figure hits and image-as-query retrieval.
9
10	## Features
11
12	- Hybrid search — Vector + full-text with Reciprocal Rank Fusion
13	- Multimodal & cross-modal search — Multimodal embedders (vLLM, VoyageAI, Cohere) put picture vectors in the same space as text; supports text-as-query → figure hits and image-as-query
14	- Question answering — RAG skill with citations (page numbers, section headings)
15	- Vision QA — Vision-capable models receive figure bytes alongside chunk text
16	- Reranking — local cross-encoders, Cohere, Zero Entropy, or vLLM
17	- Analysis skill — Complex analytical tasks via sandboxed Python code execution (aggregation, computation, multi-document analysis)
18	- Conversational RAG — Chat TUI and web application for multi-turn conversations with session memory
19	- Document structure — Stores full [DoclingDocument](https://docling-project.github.io/docling/concepts/docling_document/), enabling structure-aware context expansion
20	- Multiple providers — Embeddings: Ollama, OpenAI, VoyageAI, Cohere, LM Studio, vLLM (multimodal via `multimodal: true` on vLLM/VoyageAI/Cohere). QA: any model supported by Pydantic AI
21	- Local-first — Embedded LanceDB, no servers required. Also supports S3, GCS, Azure, and LanceDB Cloud
22	- CLI & Python API — Full functionality from command line or code
23	- MCP server — Expose as tools for AI assistants (Claude Desktop, etc.)
24	- Visual grounding — View chunks highlighted on original page images
25	- Production ingester — Long-lived `haiku-ingester` service with persistent SQLite queue, async worker pool with retries and a dead-letter queue, FS / HTTP / S3 / WebDAV source adapters, FastAPI control plane, and a browser dashboard for operators. See [docs/ingester.md](docs/ingester.md).
26	- Tags — Name database states with `haiku-rag tag` and roll back to them
27	- Inspector — TUI for browsing documents, chunks, and search results
28
29	## Installation
30
31	Python 3.12 or newer required
32
33	### Full Package (Recommended)
34
35	```bash
36	pip install haiku.rag
37	```
38
39	Includes all features: document processing, all embedding providers, and rerankers.
40
41	Using [uv](https://docs.astral.sh/uv/)? `uv pip install haiku.rag`
42
43	### Slim Package (Minimal Dependencies)
44
45	```bash
46	pip install haiku.rag-slim
47	```
48
49	Install only the extras you need. See the [Installation](https://ggozad.github.io/haiku.rag/installation/) documentation for available options.
50
51	## Quick Start
52
53	> Note: Requires an embedding provider (Ollama, OpenAI, etc.). See the [Tutorial](https://ggozad.github.io/haiku.rag/tutorial/) for setup instructions.
54
55	```bash
56	# Index a PDF
57	haiku-rag add-src paper.pdf
58
59	# Search
60	haiku-rag search "attention mechanism"
61
62	# Ask questions with citations
63	haiku-rag ask "What datasets were used for evaluation?"
64
65	# Analyze — complex analytical tasks via code execution
66	haiku-rag analyze "How many documents mention transformers?"
67
68	# Interactive chat — multi-turn conversations with memory
69	haiku-rag chat
70
71	# Continuously ingest from configured sources (FS, HTTP, S3, WebDAV)
72	haiku-ingester serve
73	```
74
75	See [Configuration](https://ggozad.github.io/haiku.rag/configuration/) for customization options.
76
77	## Python API
78
79	```python
80	from haiku.rag.client import HaikuRAG
81
82	async with HaikuRAG("knowledge.lancedb", create=True) as rag:
83	# Index documents
84	await ra

People who liked this also viewed

6 picks

Type

ddgsA metasearch library that aggregates results from diverse web search servicesMCP ServersMay 20266.5M2.8k
firecrawl-mcp-server🔥 Official Firecrawl MCP Server - Adds powerful web scraping and search to Cursor, Claude and any other LLM clients.MCP ServersJul 2026480k7.0k
fliGoogle Flights MCP, CLI and Python LibraryMCP ServersJul 2026211k3.0k
sembleFast and Accurate Code Search for Agents. Uses ~98% fewer tokens than grep+readMCP ServersJul 2026140k5.7k
duckduckgo-mcp-serverA Model Context Protocol (MCP) server that provides web search capabilities through DuckDuckGo, with additional features for content fetching and parsing.MCP ServersJul 2026121k1.4k
wikipedia-mcpA Model Context Protocol (MCP) server that retrieves information from Wikipedia to provide context to LLMs.MCP ServersMay 2026113k274

haiku.rag

5.7kinstalls/mo · pypi547stars35forks

View on GitHub

How to install haiku.rag?

ggozad/haiku.rag

$git clone https://github.com/ggozad/haiku.rag

Installs into the current project.

Summary

Agentic RAG built on LanceDB, Pydantic AI, and Docling.

Files · 1

View on GitHub

README.md

1	# Haiku RAG
2
3	[![Tests](https://github.com/ggozad/haiku.rag/actions/workflows/test.yml/badge.svg)](https://github.com/ggozad/haiku.rag/actions/workflows/test.yml)
4	[![codecov](https://codecov.io/gh/ggozad/haiku.rag/graph/badge.svg)](https://codecov.io/gh/ggozad/haiku.rag)
5
6	Agentic RAG built on [LanceDB](https://lancedb.com/), [Pydantic AI](https://ai.pydantic.dev/), and [Docling](https://docling-project.github.io/docling/).
7
8	> New: vision and multimodal search. Picture-aware ingestion captures embedded figure bytes; vision-capable QA models receive them alongside text. Multimodal embedders put picture vectors in the same space as text, enabling text-as-query → figure hits and image-as-query retrieval.
9
10	## Features
11
12	- Hybrid search — Vector + full-text with Reciprocal Rank Fusion
13	- Multimodal & cross-modal search — Multimodal embedders (vLLM, VoyageAI, Cohere) put picture vectors in the same space as text; supports text-as-query → figure hits and image-as-query
14	- Question answering — RAG skill with citations (page numbers, section headings)
15	- Vision QA — Vision-capable models receive figure bytes alongside chunk text
16	- Reranking — local cross-encoders, Cohere, Zero Entropy, or vLLM
17	- Analysis skill — Complex analytical tasks via sandboxed Python code execution (aggregation, computation, multi-document analysis)
18	- Conversational RAG — Chat TUI and web application for multi-turn conversations with session memory
19	- Document structure — Stores full [DoclingDocument](https://docling-project.github.io/docling/concepts/docling_document/), enabling structure-aware context expansion
20	- Multiple providers — Embeddings: Ollama, OpenAI, VoyageAI, Cohere, LM Studio, vLLM (multimodal via `multimodal: true` on vLLM/VoyageAI/Cohere). QA: any model supported by Pydantic AI
21	- Local-first — Embedded LanceDB, no servers required. Also supports S3, GCS, Azure, and LanceDB Cloud
22	- CLI & Python API — Full functionality from command line or code
23	- MCP server — Expose as tools for AI assistants (Claude Desktop, etc.)
24	- Visual grounding — View chunks highlighted on original page images
25	- Production ingester — Long-lived `haiku-ingester` service with persistent SQLite queue, async worker pool with retries and a dead-letter queue, FS / HTTP / S3 / WebDAV source adapters, FastAPI control plane, and a browser dashboard for operators. See [docs/ingester.md](docs/ingester.md).
26	- Tags — Name database states with `haiku-rag tag` and roll back to them
27	- Inspector — TUI for browsing documents, chunks, and search results
28
29	## Installation
30
31	Python 3.12 or newer required
32
33	### Full Package (Recommended)
34
35	```bash
36	pip install haiku.rag
37	```
38
39	Includes all features: document processing, all embedding providers, and rerankers.
40
41	Using [uv](https://docs.astral.sh/uv/)? `uv pip install haiku.rag`
42
43	### Slim Package (Minimal Dependencies)
44
45	```bash
46	pip install haiku.rag-slim
47	```
48
49	Install only the extras you need. See the [Installation](https://ggozad.github.io/haiku.rag/installation/) documentation for available options.
50
51	## Quick Start
52
53	> Note: Requires an embedding provider (Ollama, OpenAI, etc.). See the [Tutorial](https://ggozad.github.io/haiku.rag/tutorial/) for setup instructions.
54
55	```bash
56	# Index a PDF
57	haiku-rag add-src paper.pdf
58
59	# Search
60	haiku-rag search "attention mechanism"
61
62	# Ask questions with citations
63	haiku-rag ask "What datasets were used for evaluation?"
64
65	# Analyze — complex analytical tasks via code execution
66	haiku-rag analyze "How many documents mention transformers?"
67
68	# Interactive chat — multi-turn conversations with memory
69	haiku-rag chat
70
71	# Continuously ingest from configured sources (FS, HTTP, S3, WebDAV)
72	haiku-ingester serve
73	```
74
75	See [Configuration](https://ggozad.github.io/haiku.rag/configuration/) for customization options.
76
77	## Python API
78
79	```python
80	from haiku.rag.client import HaikuRAG
81
82	async with HaikuRAG("knowledge.lancedb", create=True) as rag:
83	# Index documents
84	await ra

People who liked this also viewed

6 picks

Type