Graql Instructions

Graql is the language for the Grakn knowledge graph. Whether it’s through the Graql Console or one of the Grakn Clients, Grakn accepts instructions and provides answers only in its own language - Graql.

Graql is declarative. In simple terms, when writing Graql code, you simply describe what you would like to achieve, rather than how it’s going to get there. The implementation of a Graql query is hidden under the hood and never affects you as the developer.

Graql is intuitive. Graql queries can look a lot like questions the same way they are formed in mind, as opposed to a complicated series of statements that are difficult to follow and, at scale, become confusing It’s important to note that the intuitiveness of a Graql query originates from the schema. The better the knowledge graph represents the reality of its dataset via its schema, the more intuitive writing and reading Graql queries become.

In this section, we learn how a Graql query can be written to:

  • Match specific patterns in the data and schema
  • Insert new instances of conept types into the knowledge graph
  • Get back desired data/knowledge
  • Aggregate values over a set of data
  • Delete instances of concept types from the knowledge graph
  • Compute distributed analytics over a large set of data
  • Update instances of concept types

Graql Answers

Some Graql queries such as get, aggregate and compute are expected to return an answer. Depending on the type of query, the structure of the answer may be different (more on Answer types in the Client API section). As a part of the Concept API, we can call various methods on the concepts contained in an answer to retrieve more explicit and implicit knowledge about them and their surroundings.

Reserved Keywords

The following keywords are reserved and meant to only be used by Graql in the queries.

aggregate, asc
compute, contains, count
delete, desc
get, group
id, in, insert
label, limit
match, max, mean, median, min
offset, order
std, sum


Graql is a powerful language that is easy to learn, write and read.

In the sections that follow, we learn about how we can use Graql to query and modify the data instances as well as the underlying schema.