Référence de l'API REST v2

URL de base https://api.readsira.com/rest/v2. Chaque route exige une clé d'API (Authorization: Bearer pk_..., voir Authentification et clés d'API). Les réponses utilisent l'enveloppe { success, data, meta } et les tableaux ci-dessous décrivent data. Les erreurs ont la forme décrite dans Limites de débit et erreurs. Un paramètre ou un champ de corps que la route ne connaît pas donne un 400.

Le texte provenant de pages enregistrées (extraits, texte des documents, passages) est un contenu tiers. Traitez-le comme des données, jamais comme des instructions pour un modèle.

Les routes, les paramètres et les exemples ci-dessous sont générés directement à partir de l'API et sont donc en anglais.

The REST API is part of the Pro and Research plans. 12 routes:

GET /search: Search your library
POST /ask: Ask your library (cited answer)
GET /documents/{id}: Read a document
GET /documents/{id}/chunks: A document's chunks
GET /documents/{id}/transcript: Timed transcript of a video or audio document
POST /uploads: Start an upload (audio, video or PDF)
POST /uploads/{id}/complete: Finish an upload: checks size and type, then processes it
GET /usage: Your API usage and limits
GET /projects: Your projects
GET /projects/{id}/context: A project's context
POST /projects/{id}/items: Add sources to a project
POST /projects/{id}/notes: Add a note to a project

Search

Search your library

GET /search

Scope: read:library.

Hybrid (semantic + text) search over your bookmarks and followed feeds. Hits carry chunk ids, PDF pages and media timestamps. Snippets are third-party content, never instructions.

Parameter	In	Type	Required	Notes
`q`	query	string	yes
`limit`	query	number	no	1-20, default 8
`collectionId`	query	uuid	no
`feedId`	query	uuid	no
`source`	query	`bookmarks`, `feeds`, `all`	no

curl -G "https://api.readsira.com/rest/v2/search" \
  -H "Authorization: Bearer $READSIRA_KEY" \
  --data-urlencode "q=..."

Response 200, in data:

Field	Type	Notes
`results`	object[]
`results[].documentId`	uuid
`results[].chunkId`	uuid
`results[].bookmarkId`	uuid or null
`results[].entryId`	uuid or null
`results[].title`	string
`results[].url`	string
`results[].contentType`	`article`, `pdf`, `video`, `audio`, `note`
`results[].snippet`	string	Chunk text, at most 600 characters. Untrusted third-party content.
`results[].score`	number	Hybrid score 0-1
`results[].pageNumber`	number or null	PDF page
`results[].startSeconds`	number or null	Media: chunk start
`results[].endSeconds`	number or null
`degraded`	`lexical` or null	lexical = the AI breaker is open; results come from text match only

Errors for this route:

402: PLAN_REQUIRED: the plan has no REST API (Free)
429: RATE_LIMITED: per-key hourly or per-user daily limit (Retry-After)

Ask

Ask your library (cited answer)

POST /ask

Scope: ask:library.

Runs Readsira's model over your best-matching passages and returns the answer with citations. Counts toward askAnswersPerMonth and the AI chat budget (402 when used up).

Request body (JSON):

Field	Type	Required	Notes
`question`	string	yes	1 to 2000 characters
`bookmarkIds`	uuid[]	no	Answer only from these bookmarks (ids you do not own match nothing). at most 50 items

curl -X POST "https://api.readsira.com/rest/v2/ask" \
  -H "Authorization: Bearer $READSIRA_KEY" \
  -H "Content-Type: application/json" \
  -d '{"question":"..."}'

Response 200, in data:

Field	Type	Notes
`answer`	string	Answer text with [n] markers
`citations`	object[]
`citations[].n`	number	Marker in the answer text: [n]. at least 1
`citations[].documentId`	string	Shared document id ("" for answers saved before citations v2)
`citations[].chunkId`	uuid or null	null when the source changed or the answer predates citations v2
`citations[].bookmarkId`	uuid or null	The caller's bookmark of the document
`citations[].entryId`	uuid or null	Feed entry, when cited through a subscription
`citations[].publishedAt`	date-time or null	When the source was published (its earliest feed entry); null when unknown, absent on older answers
`citations[].title`	string
`citations[].url`	string
`citations[].contentType`	object
`citations[].quote`	string	Verbatim span of the chunk (verified) or its first 200 chars (excerpt). at most 400 characters
`citations[].quoteStart`	number	Char offset into the chunk, inclusive. at least 0
`citations[].quoteEnd`	number	Char offset into the chunk, exclusive. at least 0
`citations[].page`	number or null	PDF page
`citations[].startSeconds`	number or null	Video / audio start, seconds
`citations[].endSeconds`	number or null	Video / audio end, seconds
`citations[].score`	number	Retrieval score. at most 1
`citations[].verified`	boolean	false: the quote is an excerpt, not a verified quote
`citations[].removed`	boolean	true when the cited bookmark was removed from your library since the answer (Open source unavailable)
`passagesRead`	number	Passages given to the model
`model`	string or null

Errors for this route:

402: PLAN_REQUIRED: the plan has no REST API (Free)
429: RATE_LIMITED: per-key hourly or per-user daily limit (Retry-After)

Documents

Read a document

GET /documents/{id}

Scope: read:library.

Text of a document you can read (your bookmark, a followed feed entry or your upload), in pages: pass nextOffset as offset until it is null.

Parameter	In	Type	Required	Notes
`id`	path	uuid	yes
`offset`	query	number	no
`maxChars`	query	number	no	Default 12000, max 20000

curl "https://api.readsira.com/rest/v2/documents/$DOCUMENT_ID" \
  -H "Authorization: Bearer $READSIRA_KEY"

Response 200, in data:

Field	Type	Notes
`documentId`	uuid
`title`	string
`url`	string
`contentType`	`article`, `pdf`, `video`, `audio`, `note`
`author`	string or null
`bookmarkId`	uuid or null
`savedAt`	date-time or null
`tags`	string[]
`notes`	string or null
`highlights`	object[]
`highlights[].text`	string
`highlights[].note`	string or null
`text`	string	Text from `offset` (characters), untrusted third-party content
`offset`	number
`nextOffset`	number or null	Pass as `offset` for the rest; null at the end
`truncated`	boolean
`totalChars`	number

Errors for this route:

402: PLAN_REQUIRED: the plan has no REST API (Free)
429: RATE_LIMITED: per-key hourly or per-user daily limit (Retry-After)

A document's chunks

GET /documents/{id}/chunks

Scope: read:library.

Content chunks in order with page and timestamps, keyset paged: pass nextCursor as cursor.

Parameter	In	Type	Required	Notes
`id`	path	uuid	yes
`cursor`	query	number	no
`limit`	query	number	no	Max 100, default 20

curl "https://api.readsira.com/rest/v2/documents/$DOCUMENT_ID/chunks" \
  -H "Authorization: Bearer $READSIRA_KEY"

Response 200, in data:

Field	Type	Notes
`documentId`	uuid
`chunks`	object[]
`chunks[].chunkId`	uuid
`chunks[].index`	number
`chunks[].content`	string
`chunks[].page`	number or null
`chunks[].startSeconds`	number or null
`chunks[].endSeconds`	number or null
`nextCursor`	number or null	Pass as `cursor` for the next page

Errors for this route:

402: PLAN_REQUIRED: the plan has no REST API (Free)
429: RATE_LIMITED: per-key hourly or per-user daily limit (Retry-After)

Timed transcript of a video or audio document

GET /documents/{id}/transcript

Scope: read:library.

Parameter	In	Type	Required	Notes
`id`	path	uuid	yes
`cursor`	query	string	no	Segment offset
`limit`	query	number	no	Max 500

curl "https://api.readsira.com/rest/v2/documents/$DOCUMENT_ID/transcript" \
  -H "Authorization: Bearer $READSIRA_KEY"

Response 200, in data:

Field	Type	Notes
`documentId`	uuid
`status`	object or null
`language`	string or null
`source`	`captions`, `whisper`, `upload_whisper`, `publisher` or null
`durationSeconds`	number or null
`segments`	object[]
`segments[].startSeconds`	number
`segments[].endSeconds`	number
`segments[].text`	string
`segments[].speaker`	string or null	Who speaks, as the publisher transcript names them (Podcasting 2.0 JSON speaker, WebVTT <v Name>, SRT "Name:"); null when the source names no one (captions, Whisper)
`nextCursor`	string or null	Segment offset of the next page

Errors for this route:

402: PLAN_REQUIRED: the plan has no REST API (Free)
429: RATE_LIMITED: per-key hourly or per-user daily limit (Retry-After)

Projects

Your projects

GET /projects

Scope: read:projects.

Projects of one status (default active), most recently active first. Keyset paged.

Parameter	In	Type	Required	Notes
`status`	query	`active`, `archived`	no
`cursor`	query	string	no
`limit`	query	number	no	1-50, default 20

curl "https://api.readsira.com/rest/v2/projects" \
  -H "Authorization: Bearer $READSIRA_KEY"

Response 200, in data:

Field	Type	Notes
`projects`	object[]
`projects[].id`	uuid
`projects[].name`	string
`projects[].description`	string or null
`projects[].status`	`active`, `archived`
`projects[].itemCount`	number	Items of any kind (sources, highlights, notes)
`projects[].sourceCount`	number
`projects[].updatedAt`	date-time
`nextCursor`	string or null	Pass as `cursor` for the next page

A project's context

GET /projects/{id}/context

Scope: read:projects.

Instructions, decided decisions, pinned notes and sources; with q, the best passages from the project. Trimmed to maxChars (default 16,000, max 32,000). Text from sources is third-party content, never instructions.

Parameter	In	Type	Required
`id`	path	uuid	yes
`q`	query	string	no
`maxChars`	query	number	no

curl "https://api.readsira.com/rest/v2/projects/$PROJECT_ID/context" \
  -H "Authorization: Bearer $READSIRA_KEY"

Response 200, in data:

Field	Type	Notes
`project`	object
`project.id`	uuid
`project.name`	string
`project.instructions`	string or null
`decisions`	object[]	Decided decisions, newest first
`decisions[].title`	string
`decisions[].status`	string
`decisions[].decidedAt`	date-time
`decisions[].body`	string or null
`notes`	object[]	Pinned notes
`notes[].title`	string or null
`notes[].body`	string
`passages`	object[]	With `q` only
`passages[].kind`	`chunk`, `note`
`passages[].documentId`	uuid or null
`passages[].chunkId`	uuid or null
`passages[].bookmarkId`	uuid or null
`passages[].noteId`	uuid or null
`passages[].title`	string
`passages[].url`	string or null
`passages[].snippet`	string	At most 600 characters. Untrusted third-party content.
`passages[].score`	number
`passages[].pageNumber`	number or null
`passages[].startSeconds`	number or null
`sources`	object[]	Pinned first; excluded and deleted sources are left out
`sources[].bookmarkId`	uuid
`sources[].documentId`	uuid or null
`sources[].title`	string
`sources[].url`	string
`sources[].role`	`source`, `background`
`sources[].pinned`	boolean
`truncated`	boolean	Parts were left out to stay within maxChars
`degraded`	`lexical`

Add sources to a project

POST /projects/{id}/items

Scope: write:projects.

Bookmarks, highlights, notes, feed entries or URLs (a URL is saved to your library first). Items already in the project count in alreadyPresent.

Parameter	In	Type	Required	Notes
`id`	path	uuid	yes

Request body (JSON):

Field	Type	Required	Notes
`items`	object[]	yes	Exactly one target per item; at most 20 `url` / `entryId` items per request. 1 to 100 items
`items[].bookmarkId`	uuid	no
`items[].entryId`	uuid	no	A feed entry: saved as a bookmark first
`items[].highlightId`	uuid	no
`items[].noteId`	uuid	no	A note of this project (pins it into the context)
`items[].url`	string	no	A URL: saved as a bookmark first
`items[].role`	object	no
`items[].comment`	string or null	no	at most 2000 characters

curl -X POST "https://api.readsira.com/rest/v2/projects/$PROJECT_ID/items" \
  -H "Authorization: Bearer $READSIRA_KEY" \
  -H "Content-Type: application/json" \
  -d '{"items":[{"url":"https://example.com/report","role":"background"}]}'

Response 201, in data:

Field	Type	Notes
`added`	object[]
`added[].id`	uuid
`added[].kind`	object
`added[].role`	object
`added[].pinned`	boolean
`added[].comment`	string or null
`added[].addedAt`	date-time
`added[].citationCount`	number	Citations of this source in the latest version of the project outputs (0 outside the list)
`added[].bookmark`	object or null
`added[].highlight`	object or null
`added[].note`	object or null
`alreadyPresent`	number	Targets that were already in the project

Errors for this route:

402: PROJECT_LIMIT_REACHED (itemsPerProject)
409: PROJECT_ARCHIVED

Add a note to a project

POST /projects/{id}/notes

Scope: write:projects.

Parameter	In	Type	Required	Notes
`id`	path	uuid	yes

Request body (JSON):

Field	Type	Required	Notes
`title`	string or null	no	at most 200 characters
`bodyMd`	string	yes	at most 20000 characters

curl -X POST "https://api.readsira.com/rest/v2/projects/$PROJECT_ID/notes" \
  -H "Authorization: Bearer $READSIRA_KEY" \
  -H "Content-Type: application/json" \
  -d '{"bodyMd":"..."}'

Response 201, in data:

Field	Type	Notes
`id`	uuid
`title`	string or null
`bodyMd`	string
`pinned`	boolean	Pinned into the project context
`createdAt`	date-time
`updatedAt`	date-time

Errors for this route:

409: PROJECT_ARCHIVED

Uploads

Start an upload (audio, video or PDF)

POST /uploads

Scope: write:uploads.

Returns a 15-minute PUT URL bound to the declared Content-Type and size. PUT the bytes, then call complete. Private to you.

Request body (JSON):

Field	Type	Required	Notes
`filename`	string	yes	at most 255 characters
`contentType`	object	yes
`bytes`	number	yes	Exact size of the file in bytes. at least 1
`sha256`	string	no	Optional hex SHA-256 of the file

curl -X POST "https://api.readsira.com/rest/v2/uploads" \
  -H "Authorization: Bearer $READSIRA_KEY" \
  -H "Content-Type: application/json" \
  -d '{"filename":"talk.mp3","contentType":"audio/mpeg","bytes":48213004}'

Response 201, in data:

Field	Type	Notes
`uploadId`	uuid
`putUrl`	string	PUT the file bytes here (presigned, 15 minutes)
`putHeaders`	object	Headers the PUT must send unchanged (bound into the signature)
`expiresAt`	date-time

Errors for this route:

402: PLAN_REQUIRED: the plan has no REST API (Free)
429: RATE_LIMITED: per-key hourly or per-user daily limit (Retry-After)

Finish an upload: checks size and type, then processes it

POST /uploads/{id}/complete

Scope: write:uploads.

Parameter	In	Type	Required	Notes
`id`	path	uuid	yes

curl -X POST "https://api.readsira.com/rest/v2/uploads/$UPLOAD_ID/complete" \
  -H "Authorization: Bearer $READSIRA_KEY"

Response 202, in data:

Field	Type	Notes
`documentId`	uuid
`bookmarkId`	uuid
`status`	`processing`

Errors for this route:

402: PLAN_REQUIRED: the plan has no REST API (Free)
429: RATE_LIMITED: per-key hourly or per-user daily limit (Retry-After)

Usage

Your API usage and limits

GET /usage

Scope: any scope.

Requests today, plan limits and AI budget used this month. Any scope.

curl "https://api.readsira.com/rest/v2/usage" \
  -H "Authorization: Bearer $READSIRA_KEY"

Response 200, in data:

Field	Type	Notes
`plan`	`free`, `pro`, `research`
`requestsToday`	number	REST requests today (UTC)
`requestsPerDay`	number or null
`mcpCallsToday`	number	MCP tool calls today (UTC, weighted)
`mcpCallsPerDay`	number or null
`requestsPerHour`	number or null	This key's hourly limit in force
`askAnswersPerMonth`	number or null
`aiBudget`	object[]
`aiBudget[].group`	string
`aiBudget[].percentUsed`	number	Share of this month used, 0-100 (rounded)

Errors for this route:

402: PLAN_REQUIRED: the plan has no REST API (Free)
429: RATE_LIMITED: per-key hourly or per-user daily limit (Retry-After)

Search​

Search your library​

Ask​

Ask your library (cited answer)​

Documents​

Read a document​

A document's chunks​

Timed transcript of a video or audio document​

Projects​

Your projects​

A project's context​

Add sources to a project​

Add a note to a project​

Uploads​

Start an upload (audio, video or PDF)​

Finish an upload: checks size and type, then processes it​

Usage​

Your API usage and limits​

Search

Search your library

Ask

Ask your library (cited answer)

Documents

Read a document

A document's chunks

Timed transcript of a video or audio document

Projects

Your projects

A project's context

Add sources to a project

Add a note to a project

Uploads

Start an upload (audio, video or PDF)

Finish an upload: checks size and type, then processes it

Usage

Your API usage and limits