Product

What Bolt Search does.

A hosted full-text search service. Send documents over a REST API or sync them from PostgreSQL, search them with an Elasticsearch-compatible API, and manage it all from a web console.

Indexes and field mappings

What it does: An index is a collection of documents with a schema. You declare each field’s type when you create the index, and set optional stored, indexed and fast flags.
How you use it: PUT /v1/{index} with a mappings object, or Create Index in the web console. Without mappings, a PostgreSQL sync job creates the index and infers the types. Mappings cannot be changed afterwards: to change them, create a new index and reindex.
What you see: GET /v1/{index} returns the mappings and the document count, and GET /v1/{index}/_mapping returns the mappings alone. The console’s Indices page lists each index with its health and document count.

Type	Aliases	How it is stored and searched	Typical use
`text`	—	Full-text search with BM25 scoring (tokenized).	Titles, descriptions, article bodies
`keyword`	`string`	Exact match and aggregations; not tokenized. Set "fast": true to aggregate.	Categories, tags, status codes
`integer`	`long`, `i64`	64-bit signed integer.	Counts, years, quantities
`float`	`double`, `f64`	64-bit floating point.	Prices, ratings
`boolean`	`bool`	True or false; range queries work on it too.	Flags such as in_stock
`date`	`datetime`	RFC 3339 timestamp, for example 2026-01-15T10:30:00Z.	Created and updated times

Documents and bulk indexing

What it does: Store JSON documents under an id of your choice. Indexing an existing id replaces the whole document; there is no partial update, so send the complete document each time.
How you use it: PUT /v1/{index}/_doc/{id} for one document; POST /v1/_bulk with NDJSON for many index and delete actions at once; POST /v1/{index}/_delete_by_query to remove everything that matches a query.
What you see: Index responses carry result and _version. A bulk response lists the status of every action and an errors flag.

{"index":{"_index":"movies","_id":"1"}}
{"title":"The Matrix","year":1999,"genre":"sci-fi"}
{"delete":{"_index":"movies","_id":"3"}}

Search and the query language

What it does: Full-text search with BM25 scoring, exact filters, boolean combinations and ranges, plus sorting, pagination and aggregations. It is a subset of the Elasticsearch query language: match_all, match, term, bool and range, with terms, metric and range aggregations.
How you use it: POST /v1/{index}/_search with a JSON body. Search several indexes at once by naming up to 20, separated by commas. _count and scroll are available too.
What you see: Hits with _score and _source, the total hit count, and an aggregations object keyed by the names you chose.

Every clause with an example is in the query reference.

PostgreSQL sync

What it does: Keeps an index in step with a PostgreSQL table, or with a SELECT you write. A job polls for rows past a saved cursor and indexes them. A scheduled full resync can rebuild the index and swap it in.
How you use it: Connect a database from Connections: paste a postgres:// URL, pick tables, and the console suggests the id and cursor columns and creates the syncs. Nothing is installed in your database; the job only runs SELECT statements.
What you see: Each sync’s page shows its status, last cursor, documents synced, error count, live log and resync history.

Sync articlesRunning

public.articles → articles

Total synced: 12,500
Last cursor: 2026-01-01T00:00:00Z
Errors: 0

07:00:01 [info] polled public.articles, batch of 1000 rows

07:00:01 [info] bulk indexed 1000 documents into articles

07:00:02 [info] full batch returned, polling again

07:00:03 [info] cursor saved

Sample job. How change tracking works.

Web console

What it does: Everything above from a browser: indexes, documents, connections, syncs, analytics, API keys, webhooks, members, allowed origins and billing, for the project you pick in the project switcher.
How you use it: Sign in to the app. The sidebar has Overview; Data: Indices and Connections; Search: Explorer and Analytics; Developers: API keys and Webhooks; and Project: Settings, where Billing is a tab. You can choose a light, dark or system theme.
What you see: The screens below show sample data.

IndicesCreate, list and delete indexes.ExplorerRun a query, read the hits.ConnectionsPostgreSQL connections and their syncs.SyncsState, cursor and logs.AnalyticsWhat people searched for.API keysKeys and roles.

Sample project/Indices

IndicesCreate index

NameHealthDocs

moviesGreen5

articlesGreen12,500

draftsYellow21

Create index asks for a name and optional field mappings: text, keyword, integer, float, boolean or date.

Create: A name and optional field mappings. Without mappings, a sync job creates the index.
List: Name, health and document count.
Types: text, keyword, integer, float, boolean, date.

Sample project/Explorer

Explorer

Index

movies

Search query

{"match":{"title":"matrix"}}JSON

Showing 1–3 of 3 results in 1 ms

IDScoreSource

12.3112{"title":"The Matrix","year":1999,"genre":"sci-fi"}

21.7405{"title":"The Matrix Reloaded","year":2003,"genre":"sci-fi"}

31.7405{"title":"The Matrix Revolutions","year":2003,"genre":"sci-fi"}

Input: Plain text or a JSON query clause such as {"match":{"title":"hello"}}, not a full request body.
Blank: match_all: browse every document.
Columns: ID, Score and Source, plus how many results and how long the search took.
Metering: Every search is a query attempt that uses credits, including one with no results.

Sample project/Connections

ConnectionsConnect database

app on db.example.comdb.example.com:5432/app

Sync articlespublic.articles → articlesRunning

Sync eventspublic.events → eventsStopped

The connection is tested before it is saved. The password is never shown again.

Connect: Paste a postgres:// URL or fill in the fields; the connection is tested before it is saved.
Tables: Pick tables; each gets a suggested index name, id column and cursor column.
Passwords: Never returned by the API.

Sample project/Connections/Sync articles

Sync articles

Sync articlesRunning

public.articles → articles

Total synced: 12,500
Last cursor: 2026-01-01T00:00:00Z
Errors: 0

07:00:01 [info] polled public.articles, batch of 1000 rows

07:00:01 [info] bulk indexed 1000 documents into articles

07:00:02 [info] full batch returned, polling again

07:00:03 [info] cursor saved

Actions: Start, stop, reset cursor and full resync.
Live log: Each sync’s page shows its log as it runs.
Resyncs: History of each full resync with its outcome.

Sample project/Analytics

Analytics

movies

24h7d30d

Top queriesZero resultsLow CTR

QueryCountAvg results

matrix423

{"term":{"genre":"sci-fi"}}3112

inception181

dark knight92

Scope: One index and a time range: 24 hours, 7 days or 30 days.
Tabs: Top Queries, Zero Results and Low CTR.

Sample project/API keys

API keys

Backend servicebsk_a1b2c3…editorActive

Search widgetbsk_d4e5f6…viewerActive

Old CI keybsk_0a9b8c…editorRevoked

New keys are shown only once. Copy and store them securely.

API keys: One project and a role, editor or viewer. The full key is shown once; the list shows prefixes.
Settings: General, Members, Allowed origins and Billing tabs.

Projects, API keys and roles

What it does: A project is the unit of isolation: its own indexes, data sources, sync jobs, members, API keys and credit balance. Signing up creates your account and a project you own.
How you use it: Create keys in Settings, API Keys: give the key a name and a role, editor or viewer, and copy it when it is shown, because it is shown only once. Settings also has Members and Allowed Origins, which controls which websites may call the project from a browser.
What you see: Keys are listed by prefix, such as bsk_a1b2c3, with their role and status. A revoked key stops working at once.

What it allows	viewer	editor	owner
Search, get documents and indexes, list indexes	Yes	Yes	Yes
View data sources and sync jobs	Yes	Yes	Yes
Index and delete documents, create and delete indexes, bulk	—	Yes	Yes
Manage data sources and sync jobs	—	Yes	Yes
Manage project members, API keys and settings	—	—	Yes

Search analytics

What it does: Shows what people search for in an index: the most frequent queries, the queries that returned nothing, and the queries with a low click-through rate.
How you use it: Open Analytics in the console, choose an index and a range of 24 hours, 7 days or 30 days. The same data is available as GET /v1/{index}/_analytics/…. Report a click with POST /v1/{index}/_track to feed click-through rates.
What you see: Three tabs: Top Queries, Zero Results and Low CTR. Each search records its text, result count and latency. See Security for how long this data is kept.

Usage and credits

What it does: You pay for what you search. Authorized search, count and scroll requests are metered at 10 attempts per credit, and an _msearch request counts one attempt per contained search; writes, indexing, PostgreSQL sync and analytics are free. Credits are prepaid, in packs, and never expire.
How you use it: Open Billing in the project, see the available credits and the month’s usage, and buy a pack. The pricing page has an estimator.
What you see: A project with no credits gets HTTP 402 on metered requests until you add credits. The TypeScript SDK raises it as a BoltSearchError with statusCode 402.

Create a project, make an API key and send your first search.

Get started See pricing