I think somesoni2 has a better option with a syslog-ng relay, but you can use the built in syslog routing queue to send data from the Meraki host from Splunk out to S3; look into the _SYSLOG_ROUTING value for DEST_KEY in transforms. Then configure an outputs.conf for your S3 host.
There may be workaround, but if you're not going to use data in Splunk after indexing it, why even involve Splunk? Amazon S3 has several CLI tools (http://aws.amazon.com/cli/) which you can use to send data/files to S3 storage.
I believe it would be better to just setup a syslog-ng on a server to write Meraki logs to file and use S3 CLI tools to upload those files to s3 buckets. This way you can keep the files in exact same format they are received from Meraki.