Before anything else, the correction that saves libraries a wasted procurement cycle: Koha does not talk to RFID hardware directly. There is no RFID module and no driver.
RFID readers and their middleware present an item identifier to Koha either as keyboard input — exactly as a barcode scanner would — or over SIP2. Koha sees an item barcode. It does not know or care that a radio was involved.
This matters because it reframes the buying decision. You are not looking for RFID that "supports Koha". You are looking for RFID whose middleware speaks SIP2 or emulates a keyboard — which is nearly all of it.
How RFID actually integrates
| Layer | What it is | Talks to |
|---|---|---|
| 1. Tag | Adhesive chip inside the item, carrying its barcode | The reader, by radio |
| 2. Reader + middleware | Pad, kiosk, gate or wand, plus vendor software | Koha, via SIP2 or keyboard input |
| 3. Koha | Applies circulation rules, records the transaction | Your staff and patrons |
The integration question is entirely about layer 2. Ask any vendor one thing: does your middleware speak SIP2?
What RFID gives a library
- Faster circulation — several items read at once instead of one barcode at a time.
- Inventory in the stacks — a wand reads a shelf in seconds. Usually the largest real saving, and rarely the reason people buy.
- Security gates — the same tag does theft detection, replacing separate strips.
- Self-service — patrons issue their own items.
Honestly: the circulation speed gain is real but smaller than vendors suggest, because scanning is not usually the slow part of a desk transaction. The inventory gain is larger than most libraries expect.
Components you need
Tags
HF (13.56 MHz) is the library standard — short read range, reliable, works on stacked items. UHF reads further and is more affected by metal and liquids; some libraries use it for stock control.
The tag holds the item barcode, not patron data. Patron information must never be on a tag — anyone with a reader can read it in a corridor.
ISO 28560 governs how data is written to the tag, and it decides whether you can ever change supplier. A proprietary data model means a future vendor cannot read your existing tags and you re-tag the whole collection to switch. Insist on a standards-compliant model, in writing, before you buy.
Staff pads, self-check units, gates, wands
A pad at each desk for staff circulation; self-check units for patrons; gates at exits; a wand for inventory. Not all libraries need all four — the wand is the most commonly skipped and often the most valuable.
Tagging an existing collection
The part libraries underestimate most. Every item must be opened, tagged and linked.
A realistic rate is 100–200 items per person per hour once a rhythm is established. For 40,000 items that is roughly 250 person-hours — six to eight weeks for one person, or a fortnight for a small team.
Phase it: tag high-circulation stock first so the benefit starts early, and keep barcodes on items and readable throughout. Barcodes remain your fallback when a tag fails, and tags do fail.
What RFID costs
| Item | Indicative USD |
|---|---|
| HF tags | 0.15–0.50 each |
| Staff pad + software | 800–2,500 |
| Self-check unit | 4,000–12,000 |
| Security gate pair | 6,000–15,000 |
| Inventory wand | 2,000–5,000 |
| Tagging labour | Your staff time, or contracted |
Worked example — 40,000-item library, one branch: tags 6,000–20,000; two staff pads 1,600–5,000; one self-check 4,000–12,000; one wand 2,000–5,000; one gate pair 6,000–15,000. Total roughly USD 20,000–57,000, plus around 250 hours of tagging.
Indicative ranges, reviewed August 2026. Get written quotes. See also our cost breakdown.
RFID and Koha inventory
A wand collects barcodes from a shelf; those are reconciled against Koha's holdings to find items missing, mis-shelved or not marked as on loan. An annual stocktake that took a fortnight becomes a few days.
This is where the payback usually is. If the business case rests only on faster checkout, look again at inventory.
Problems to plan for
- Tag collision — a tall stack read at once may miss items. Staff need to verify counts.
- Metal and media — CDs and DVDs need specialised tags.
- Gate false alarms — badly tuned gates train staff to ignore them, which defeats the point.
- Middleware pinned to one Koha version — ask about upgrade compatibility before buying.
- Proprietary tag data — the lock-in described above. The most expensive mistake available here.
Do you actually need RFID?
An honest gate, because this is a large purchase.
RFID rarely pays back below roughly 20,000–30,000 circulations a year. Below that, a good barcode workflow beats badly-implemented RFID, and the money buys more stock or more staff hours.
There is also a cheaper route to much of the same benefit. Circulation from a phone camera removes the per-station scanner entirely, lets staff work in the stacks instead of hauling books to a desk, and needs no middleware server running alongside it — see running circulation from a phone. For many small and mid-size libraries that is the better purchase, and it can be scanned with a phone camera on a system we host for you, with no server to run either.
If you do go RFID, SIP2 is what makes the major hardware brands work without an integration project.
Planning a rollout
- Survey — collection size, media types, branches, exits
- Business case — where the saving actually is, honestly
- Procurement — insist on standards-compliant tag data and SIP2 middleware
- Pilot — one section, one pad, prove the workflow
- Tagging — phased, high-circulation stock first
- Integration and testing — every device, before go-live
- Staff training, including what to do when a tag fails
Planning an RFID project? We integrate RFID with Koha over SIP2 whatever brand you buy, and we will tell you honestly if the numbers do not justify it.