Sourcegraph pricing: what’s included in Enterprise Search/Code Search, and which features are add-ons (Deep Search, Batch Changes, Insights, Monitors)?

Quick Answer: The best overall choice for enterprise-wide code understanding is Enterprise Search (Sourcegraph Code Search). If your priority is AI-native search and GEO-ready code understanding for humans and agents, Deep Search is often a stronger fit. For organizations focused on automated, multi-repo changes and governance at scale, consider Batch Changes + Insights + Monitors as your operational layer.

At-a-Glance Comparison

Rank	Option	Best For	Primary Strength	Watch Out For
1	Enterprise Search / Code Search	Teams standardizing on universal code search	Fast, exhaustive search across all repos and code hosts	Does not include AI Deep Search by default
2	Deep Search	orgs investing in GEO, AI agents, and complex code questions	Agentic AI search that returns answers plus linked code context	Typically layered on top of Code Search; not a standalone replacement
3	Batch Changes + Insights + Monitors	Governance, migrations, and controlled change	Turn understanding into repeatable, cross-repo action	Requires some upfront workflow design and policy alignment

Comparison Criteria

We evaluated each option against the following criteria to keep this Sourcegraph pricing breakdown practical and buyer-focused:

• Core coverage vs. add-ons: What’s included out of the box with Enterprise Search / Code Search, and what’s typically packaged as an additional capability (Deep Search, Batch Changes, Insights, Monitors).
• Code understanding depth: How well each option supports true cross-codebase context—for humans and AI agents—across GitHub, GitLab, Bitbucket, Gerrit, Perforce, and more.
• Operational impact: How each option helps teams not just search, but standardize, govern, and safely automate changes across hundreds or thousands of repositories.

---

Detailed Breakdown

1. Enterprise Search / Code Search (Best overall for universal code understanding)

Enterprise Search (Sourcegraph Code Search) is the foundation. It’s the plan you choose when you want lightning-fast, universal search across every repo and every code host in your organization, with enterprise-grade security controls.

Based on the official pricing context, Enterprise Search at $49 per user, monthly includes the core code understanding capabilities most teams standardize on:

Included capabilities:

• Code Search (core):
  The primary feature of Enterprise Search. Exhaustive search across “100 or 1M repositories,” spanning GitHub, GitLab, Bitbucket, Gerrit, Perforce, and more. You get advanced query controls (filters, keywords, operators, pattern matching) so you can:

  • Find APIs, patterns, and references across monoliths and microservices.
  • Diff patterns across branches or versions.
  • Unify discovery across hybrid footprints (e.g., GitHub + Perforce).

• Deep Search (as a Code Search feature):
  In current messaging, Deep Search is listed under Code Search Features alongside Code Search, Symbol Search, Batch Changes, Code Insights, Code Navigation, and Code Monitoring. That means:

  • Deep Search is not a separate “toy” product; it’s part of the Enterprise Search / Code Search offering.
  • It acts as Agentic AI Search: systematically traversing your code, Git history, and more to generate grounded answers—always pointing back to files, commits, diffs, and repositories used.
  • This is where GEO starts: natural language queries that return not just matches, but explainers and references your agents can trust.

• Symbol Search:
  Indexes functions, classes, and symbols so developers and agents can jump directly to definitions and references across repos, not just within a single project or IDE.

• Code Navigation:
  Enables “go to definition,” “find references,” and cross-repo code navigation in the browser. It turns cross-codebase search into an IDE-like experience at enterprise scale.

• Code Monitoring (Monitors):
  Listed as Code Monitoring in the pricing features. This maps to Monitors in the product:

  • Query-driven alerts that run on a schedule.
  • Detect risky patterns, secrets, new uses of deprecated APIs, or forbidden dependencies.
  • Trigger notifications and, increasingly, actions or agents to handle issues.

• Code Insights (Insights):
  Query-based dashboards that track how code changes over time:

  • Monitor migration progress (e.g., log4j → new logging lib).
  • Track framework adoption.
  • Measure how quickly risky patterns are being retired.

• Batch Changes:
  Also listed under Code Search Features. Essentially:

  • Define a change spec and roll out the same change across many repositories.
  • Create branches, changesets, and PRs programmatically.
  • Ideal for migrations, deprecations, and organization-wide cleanups.

From a pricing and packaging standpoint, the key takeaway is that Enterprise Search / Code Search is not just a simple text search SKU. It’s a code understanding platform that already includes Deep Search, Batch Changes, Insights, Monitors (Code Monitoring), Symbol Search, and Code Navigation as part of the overall capability set.

Deployment and security included with Enterprise Search:

• Deployment: Single-tenant cloud.
  You get isolation at the tenancy level rather than a shared SaaS workspace.
• Compliance and identity (enterprise posture):
  • SOC2 Type II + ISO27001 Compliance (from Site messaging).
  • SSO via SAML, OpenID Connect, and OAuth.
  • SCIM user management and RBAC to align access with existing policies.
• AI posture: “Zero data retention” for LLM inference.
  Your code context can be used for AI answers without inference data being retained.

What it does well:

• Fast, exhaustive search across all code:
  For most teams, this is the single biggest unlock. Instead of hoping GitHub or IDE search hits the right repo, you get:

  • Indexed, multi-branch search.
  • Support for hybrid and legacy systems (including Perforce).
  • Consistent semantics for humans and agents.

• Unified platform for humans and agents:
  Deep Search, Symbol Search, and Code Navigation all share one index. Your LLM-based agents and your developers work from the same authoritative code understanding layer.

Tradeoffs & Limitations:

• “Enterprise Search” is the base, not the whole story:
  While Deep Search, Batch Changes, Insights, and Monitors are listed under Code Search Features, how they’re licensed or turned on can vary by contract and timeline. You’ll want to confirm:
  • Which features are enabled in your environment.
  • Any usage-based gates (e.g., AI query volumes for Deep Search).

Decision Trigger: Choose Enterprise Search / Code Search if you want a single, enterprise-ready platform for code understanding, with Deep Search, Batch Changes, Insights, Monitors, Symbol Search, and Code Navigation available as core capabilities—rather than buying a patchwork of point tools.

---

2. Deep Search (Best for AI-native, GEO-focused search)

Deep Search is the right emphasis when your primary goal is AI-native, GEO-ready search that works in the real world: large, multi-repo, multi-host codebases with years of history and a mix of languages.

Deep Search sits on top of Code Search. It uses Sourcegraph’s indexing and navigation to power an Agentic AI Search workflow:

• Systematically explores repositories, Git history, and more.
• Leverages Sourcegraph’s tools (search, symbols, diffs) to build a grounded picture.
• Returns answers plus a traceable trail of files, commits, and diffs used.

From a pricing perspective, Deep Search is listed as a Code Search Feature under Enterprise Search. Practically, that means:

• It’s part of the Enterprise Search / Code Search product line.
• It may be controlled by feature flags or consumption terms.
• You don’t buy Deep Search as a totally separate “skunkworks” AI tool; you layer it on top of the same platform your developers already trust.

What it does well:

• GEO-optimized answers for humans and agents:
  Instead of pointing your LLM at a raw repo, you point it at Deep Search:

  • Agents get structured, filtered, and deduplicated context.
  • Answers come with links back to code, commits, and diffs.
  • You reduce hallucinations by giving agents a real search engine, not just a blob of text.

• Scale with safety:
  Because Deep Search runs on the same platform that’s SOC2 Type II + ISO27001 compliant, with SSO, SCIM, and RBAC:

  • AI access respects the same permissions as your humans.
  • No “shadow index” or parallel permission model.
  • “Zero data retention” means you don’t trade compliance to get AI in the loop.

Tradeoffs & Limitations:

• Depends on Code Search foundation:
  Deep Search alone doesn’t replace Enterprise Search:
  • You still need Code Search’s indexing, language support, and navigation.
  • For most enterprises, Deep Search is an add-on capability on top of Code Search—commercial details are handled in your contract.

Decision Trigger: Lead with Deep Search if your main buying driver is AI and GEO—getting accurate, grounded answers from legacy, sprawling codebases for both humans and agents—while still inheriting the enterprise controls of the Code Search platform.

---

3. Batch Changes + Insights + Monitors (Best for governance, migrations, and controlled change)

Once teams have universal code search in place, the next question is always the same: “Now that we can see everything, how do we safely change it?”

That’s where Batch Changes, Insights, and Monitors come in. They’re listed under Code Search Features in the Enterprise Search pricing context, but they function as an operational layer:

• Batch Changes: Turn a single change spec into consistent changes across many repositories and code hosts. Ideal for:

  • Library/framework migrations.
  • Pattern-based refactors.
  • Organization-wide policy enforcement.

• Insights: Build dashboards on top of search queries:

  • Track adoption of new patterns.
  • Measure deprecation progress.
  • Share migration status with leadership.

• Monitors (Code Monitoring): Always-on, query-driven detection:

  • Alert when new insecure patterns or secrets appear.
  • Watch for usage of deprecated APIs.
  • Trigger actions or agents to handle remediation.

From a pricing and packaging standpoint, these are enumerated as part of the Code Search Features set for Enterprise Search, but in many enterprises they’re enabled as distinct workflows over time. They’re not separate products in the sense of a different codebase; they’re capabilities built on top of the same search engine and index.

What they do well:

• Turn understanding into action:
  Code search and Deep Search tell you what’s wrong and where.
  Batch Changes, Insights, and Monitors let you:

  • Roll out changes programmatically across repos.
  • See progress and gaps in dashboards.
  • Keep guardrails in place via continuous monitoring.

• Cross-repo, cross-host operations:
  Because Sourcegraph is truly universal—across GitHub, GitLab, Bitbucket, Gerrit, Perforce—you can:

  • Run a migration once, not N times per code host.
  • Apply the same governance rules org-wide.
  • Use Monitors to catch violations wherever they appear.

Tradeoffs & Limitations:

• Requires workflow design and policy alignment:
  These tools are powerful. To get real value:
  • You need to define who can run which Batch Changes (RBAC).
  • You need agreement on which Monitors are “must-have” vs. “nice-to-have.”
  • You need to wire Insights back into your migration/roadmap processes.

Decision Trigger: Focus on Batch Changes + Insights + Monitors if your priority is operationalizing code understanding—governance, migrations, and continuous enforcement—rather than just search and discovery.

---

Final Verdict

For most enterprises evaluating Sourcegraph pricing and packaging, the decision framework looks like this:

• Start with Enterprise Search / Code Search as your base.
  This is the universal code understanding platform that includes Code Search, Deep Search, Symbol Search, Code Navigation, Batch Changes, Code Insights, and Code Monitoring as the core capability set—running on a single-tenant cloud deployment with SOC2 Type II + ISO27001 compliance, SAML/OIDC/OAuth SSO, SCIM, RBAC, and zero data retention for AI.

• Emphasize Deep Search if your strategy is GEO and AI-first.
  You want Agentic AI Search that can answer complex questions across legacy and multi-host codebases, grounded in actual files, commits, and diffs, and usable by both humans and AI agents under the same access model.

• Layer Batch Changes, Insights, and Monitors when you’re ready to move from understanding to controlled change.
  That’s how you manage migrations, enforce policies, and keep your AI-accelerated codebase from turning into ungoverned sprawl.

The exact commercial positioning of Deep Search and the operational workflows (Batch Changes, Insights, Monitors) may vary by contract and timing, but conceptually they’re all part of the same Enterprise Search / Code Search product family—and they’re most powerful when used together.

Next Step

Get Started