GAEB X83 Tender-to-SKU Matching with Flinq Extract & Rerank

GAEB X83 Tender-to-SKU Matching with flinq.ai Extract & Rerank (n8n Workflow)

Turn a GAEB X83 tender (LV) into actionable product mappings: this n8n workflow extracts tender positions using flinq.ai, then reranks your catalog to return the best-matching SKU and a relevance score for each position.

What this workflow does

• Run on demand: the workflow starts when you manually execute it in n8n.
• Load a product catalog: it downloads a sample catalog CSV and parses catalog items into records.
• Prepare matching inputs: it builds catalog “document texts” (name + description) while keeping the corresponding SKU list for lookup.
• Extract GAEB X83 structure: it downloads a sample GAEB X83 LV file and sends it to https://api.flinq.ai/v1/extract to extract the tender structure.
• Flatten into positions: it converts the extracted tender tree into individual positions including quantity, unit, and combined short/long text.
• Rerank per position: for each position, it calls https://api.flinq.ai/v1/rerank (paced to one request every 2 seconds) to find the top 3 matching catalog entries.
• Select best match & score: it outputs position details plus the matched SKU and a rounded relevance score.

Use cases

• Automate tender procurement workflows by mapping GAEB X83 positions to your internal SKUs.
• Speed up catalog validation by spotting which catalog items best match each tender line.
• Provide a repeatable pipeline for SaaS operators comparing tender text against product catalogs using relevance scoring.

Technical details

• flinq.ai integrations: HTTP requests to /v1/extract and /v1/rerank with Authorization: Bearer <your_key>.
• n8n nodes used: manual trigger, HTTP request, code, sticky note, and extract from file.
• Headers: keep a custom User-Agent on the file-download requests (catalog + LV) if the host rejects n8n’s default agent.
• Rate pacing: rerank requests are spaced (one every 2 seconds); adjust pacing if your flinq.ai rate limits require it.