Sample comments behind an analysis
Requires comments:read, plus analyses:read (or analyses:run for
an an_… analysis created by the same key). A sample of comments
behind this analysis — the same mix-bar filters as
GET /posts/{postId}/analysis/comments. Hidden comments are
excluded.
Siftsy is an analysis platform, not a comment exporter. Do not use
this route to dump full comment sections. Prefer source=notables,
content, or all.
For analyses created with POST /analyses, the analysis’ own
filters and search matches are applied first, so the sample is
drawn from the comments the analysis numbers describe; the query
filters below narrow further. Returns 409 analysis_not_ready
until status is ready.
Sampled rows include sampleReasons. meta.groups lists content
topic and message membership; when more than one post is included,
each group includes postId.
Analyses with more than 1000 posts return 400.
Rate class analysis_comments (20 / minute, 300 / day per key).
Every page re-reads the comments of every post on the analysis, so
take one sample per analysis rather than paging through it.
Authentication
Bearer authentication of the form Bearer <token>, where token is your auth token.
Path parameters
Query parameters
Comparisons only: sample one side (1-based, in compare order).
Omit to sample across every side. 400 on other analyses.
How to sample comments behind this analysis. Prefer notables
(engagement + relevance), content (grouped onto creator
topics/messages), or all (union). This is evidence, not a
full-thread export.
Only comments whose overall sentiment falls in these mix-bar buckets.
Comma-separate or repeat to match several (positive,lean_positive).
Values: positive, lean_positive, neutral, lean_negative,
negative. Score version 3 uses a five-bucket split (neutral is
4.5–5.5); v1/v2 have no unused middle band.
Only comments whose content sentiment (alignment with the post /
creator) falls in these buckets. Same keys as sentiment. Score
version 2+ comments; unscored comments are excluded.
Only comments whose topic sentiment (reception of the subject)
falls in these buckets. Same keys as sentiment.
Only comments in these relevance mix-bar buckets. Canonical keys
match metrics (positive, lean_positive, neutral,
lean_negative, negative). Dashboard labels also work: on_topic,
relevant, general, off_topic.
Only comments in these consensus mix-bar buckets. Canonical keys
match metrics. Dashboard labels also work: agreeing, discussing,
debating, fighting. Consensus is how the commenter is reacting to
other commenters, not to the post.
