Home / Companies / GitLab / Blog / Post Details
Content Deep Dive

GitLab Orbit is GA: Higher agent accuracy, lower token cost

Blog post from GitLab

Post Details
Company
Date Published
Author
Corinne Dent and Angelo Rivera and Gabriel Engel
Word Count
1,680
Company Posts That Month
25
Language
English
Hacker News Points
-
Post removed?
No
Summary

GitLab Orbit is a generally available context graph for the software development lifecycle that connects code, merge requests, work items, pipelines, vulnerabilities, deployments, and ownership data in a continuously updated, permission-aware view. It is designed to give AI agents and engineers broader operational context than code-only or RAG-based approaches, reducing repeated searches and enabling questions about change impact, vulnerability exposure, CI failures, and migration dependencies across projects. GitLab reports that Orbit-connected agents required up to 45 times fewer retries and 4.5 times fewer tokens in internal testing, while Compare the Market found its Orbit-grounded code reviewer matched roughly 70% of expected inline comments across 79 merge requests, compared with 58% for RAG. Orbit can be queried through GitLab Duo, MCP-compatible tools such as Claude Code and Codex, custom REST integrations, and GitLab’s Data Explorer, and it indexes lifecycle activity independently of GitLab’s core service. Available to Premium and Ultimate customers initially on GitLab.com, with Self-Managed and Dedicated support planned, Orbit is priced through shared organizational query packs costing $30 in GitLab Credits per user monthly for 1,000 queries, with overage charges and a 30-day trial available through account teams.

Trends Found in this Post
Trend Post Mentions Total Month Mentions Posts Companies MoM
MCP 3 11,896 1,225 232 +12%
RAG 3 873 223 83 +1%
Data Pipeline 1 401 160 72 +47%
Use This Data

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.