The active proposal is documented in our list archive.

If you'd like to discuss this roadmap, please join the conversation on our user or dev list: https://tika.apache.org/mail-lists.html

Oct 2024 -- Release 3.0.0
Oct 2024 -- after the 3.0.0 release, move main branch to 4.x (and Java 17) and move tika-pipes to its own module or create a standalone tika-pipes project
??? -- start 4.0.0-BETA releases ASAP
April 2025 (6 months from 3.x release) -- end support for 2.x (and thereby Java 8)
January 2026 -- release 4.0.0
June 2026 (6 months from the from 4.x release) -- end support for 3.x (and thereby Java 11)

For some ideas on changes in 4.x, please join the conversation:

Tika Version	Minimum JRE	javax vs jakarta	Availability	Planned EOL	EOL
2.x	8	javax	now	April 2025	May 2025
3.x	11	jakarta	October 2024	June 2026 or 6 months after the release of 4.0.0, whichever comes first.
4.x	17	jakarta	January 2026 (estimated)
5.x	21	jakarta	???
6.x	25	jakarta	???

***Please note that all dates are in Open Source Standard Time which does not always neatly align with traditional calendars.***

Metadata changes in 4.x

In general, we're trying to ensure that we prefix all "user generated" metadata keys. There could be surprises if files overwrote existing metadata keys, like "Content-Type". The overarching issue tracking this is: .

HTML custom metadata keys are now prefixed with "html:" (TIKA-4357)
MAPI metadata prefix has changed to "mapi:" (TIKA-4360) and "mapi" has been removed from the key names.
"resourceName" has changed to "X-TIKA:resourceName" (side effect on TIKA-4410).
Unrecognized image metadata keys are now prefixed with "img:" (TIKA-4426).

Other metadata changes:

Office metadata prefix has changed from "meta" to "office" (TIKA-???).