What is search-module?

A guide to Airweave's search code, organised as reusable processing steps. It documents both an older search system and a newer system with instant, classic, and agentic search modes.

Use it when working on search requests, vector search, result ranking, generated answers, filters, streaming events, or the search API.

It helps developers understand where search behavior lives and avoid relying on instructions marked as partly outdated.

search-module — Cursor rule by airweave-ai

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods

npx agentmods add rules/airweave-ai/airweave/search-module

Clone the repo

git clone --depth 1 https://github.com/airweave-ai/airweave

Made for: Cursor.

Per session 0 Nothing until a file matches its globs; then the whole rule loads.

When invoked 7,049 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5	$0.00000	$0.07049
Opus 5	$0.00000	$0.03524
Sonnet 5	$0.00000	$0.01410
Haiku 4.5	$0.00000	$0.00705

Measured yesterday against content hash 2ce2c8df715c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

search-module scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.cursor/rules/search-module.mdc · 793 lines

How it starts

The opening of the file, as written. The whole thing — 793 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Airweave Search Rules

WARNING — PARTIALLY OUTDATED (March 2026)

This file documents the legacy V1 search module (airweave/search/), which uses an operation-based pipeline with Qdrant, Cerebras/OpenAI/Groq providers, and structured-output LLM calls.

The new V2 search lives in airweave/domains/search/ and implements a three-tier architecture:

Instant — direct vector search, no LLM

Classic — vector search + reranking + answer generation (replaces V1)

Agentic — multi-turn tool-calling agent loop with search/count/navigate/read/collect/finish tools

Key differences from V1:

Vector DB: Vespa (not Qdrant)

LLM adapters: airweave/adapters/llm/ with Together AI + Anthropic fallback (not Cerebras/Groq)

Reranker: airweave/adapters/reranker/ with Cohere

Filters: domains/search/types/filters.py — FilterableField/FilterOperator/FilterGroup (not Qdrant native filters)

Streaming: SSE events defined in domains/search/agentic/events.py (tool_call, thinking, search_results, etc.)

API: POST /collections/{id}/search/{tier} where tier is instant, classic, or agentic

Config: domains/search/config.py — model specs, token budgets, tier defaults

When working on domains/search/, prefer reading the actual code over this file. The sections below remain accurate for the legacy airweave/search/ module only.

Overview (Legacy V1)

The search module (@search/) implements a modular, pipeline-based architecture with composable operations. It aims to maintain search quality and flexibility.

Core Architecture

Operation-Based Pipeline

SearchRequest → SearchFactory → SearchContext → SearchOrchestrator → SearchResponse
                                    ↓
                            [Operations Pipeline]

Each operation:

Implements SearchOperation abstract base class
Declares dependencies explicitly
Reads/writes to shared state dictionary
Can be optional (graceful failure)
Executes asynchronously

Read the full file on GitHub · 793 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

yesterday First seen · 793 lines · 7,049 tokens per session scan A 2ce2c8df715c

Subscribe to this mod's changes

search-module is a cursor rule published in the GitHub repository airweave-ai/airweave (6,567 stars, last pushed 2mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 7,049 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Backend and APIs Data and AI LLM RAG Research agent-infrastructure ai ai-agents search

search-module

Getting it into your agent

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

Airweave Search Rules

Overview (Legacy V1)

Core Architecture

Operation-Based Pipeline

What this file has done since we first saw it

Other cursor rules, from other repositories

trigger.advanced-tasks

orchestration

vios-sqa

cursorrules

cursorrules

backend