Continue | Ry Walker Research

Key takeaways

Repositioned as "quality control for your software factory" — source-controlled AI checks that run on every pull request as GitHub status checks
Three surfaces remain — IDE extensions (VS Code, JetBrains), the open-source CLI agent (cn), and PR checks enforceable in CI
33K+ GitHub stars and ~$5.1M raised from Y Combinator and Heavybit — still seed-stage against well-funded competitors

FAQ

What is Continue?

Continue is an open-source AI coding platform centered on source-controlled AI checks that run on every pull request, powered by an open-source CLI agent, with VS Code and JetBrains extensions.

How much does Continue cost?

The open-source CLI and IDE extensions are free with your own API keys. Cloud pricing is pay-as-you-go at $3 per million tokens (Starter), $20/seat/month for teams (includes $10 credits), and custom enterprise pricing.

What is Continuous AI?

Continuous AI is Continue's approach to running AI checks and agents on every PR — markdown-defined checks in your repo that show up as GitHub status checks, like CI/CD for code standards enforcement.

How is Continue different from Cursor?

Cursor is a full AI-native IDE; Continue is an extension for your existing editor plus a CLI agent and automated PR checks for quality enforcement.

Executive Summary

Continue is an open-source AI coding platform that has repositioned from an IDE assistant into "quality control for your software factory" — source-controlled AI checks that run on every pull request, enforceable in CI and powered by the open-source Continue CLI.^[1]^[2] With 33,000+ GitHub stars as of June 2026, roughly 3.3M VS Code installs, and ~$5.1M raised from Y Combinator and Heavybit, Continue is a credible open-source player betting that automated, repo-defined PR review becomes standard.^[3]^[4]

Attribute	Value
Company	Continue Dev, Inc.
Founded	2023
Funding	~$5.1M total — $2.1M post-YC (2023) + $3M seed (Feb 2025), Heavybit + YC^[5]^[6]
GitHub Stars	33,600+ (June 2026)^[1]
License	Apache 2.0

Product Overview

Continue's flagship product as of mid-2026 is AI checks: markdown files stored in .continue/checks/ in your repo, each containing a name, description, and prompt. On every pull request, Continue evaluates the diff against each check and reports results as native GitHub status checks — green for pass, red for fail with suggested fixes.^[2]^[7] The pitch is "consistency over breadth": mechanically enforcing the standards your team writes down rather than offering broad, unpredictable AI review feedback.

The interactive tools remain: the open-source CLI agent (cn) — the same agent that powers the IDE extensions — and VS Code/JetBrains extensions for chat, edit, autocomplete, and agent mode.^[8]

Key Capabilities

Capability	Description
AI Checks	Markdown-defined checks run on every PR as GitHub status checks, enforceable in CI
Agent Mode	Agentic coding assistant in IDE and CLI
Chat	Ask questions and clarify code sections
Edit	Inline code modifications without leaving files
Autocomplete	Inline suggestions as you type

Product Surfaces

Surface	Description	Availability
AI Checks	PR status checks defined in-repo, GitHub integration	GA
CLI (TUI Mode)	Interactive terminal agent (`cn` command)	GA
CLI (Headless Mode)	`cn -p "prompt"` for scripts and CI/CD	GA
VS Code Extension	Chat, edit, autocomplete, agent (~3.3M installs)	GA
JetBrains Extension	Same features for IntelliJ ecosystem	GA

Technical Architecture

Continue runs locally as IDE extensions or CLI, connecting directly to your chosen LLM provider.^[7] AI checks run as a GitHub integration, evaluating PR diffs against the markdown check definitions stored in the repository and posting results as status checks.

Installation:

# macOS / Linux
curl -fsSL https://raw.githubusercontent.com/continuedev/continue/main/extensions/cli/scripts/install.sh | bash

# Then run
cn  # TUI mode for interactive development
cn -p "Generate commit message"  # Headless mode for automation

Key Technical Details

Aspect	Detail
Deployment	Local (IDE/CLI) + Cloud (AI checks on PRs)
Model(s)	BYOK for any provider, or buy credits for frontier models
Integrations	VS Code, JetBrains, GitHub, Slack, Sentry, Snyk^[9]
Open Source	Yes (Apache 2.0)

TUI Mode: For complex development tasks requiring human oversight — refactors, feature implementation, debugging. Supports @ file references, slash commands, and cn --resume for continuing sessions.^[8]

Headless Mode: For proven workflows you've tested — cn -p "your prompt" runs to completion and prints to stdout, suitable for CI/CD, git hooks, and scheduled tasks.

Strengths

Source-controlled checks — Review standards live in the repo as markdown, versioned and reviewable like code
Native GitHub integration — Checks surface as status checks, enforceable as merge gates in CI
Model-agnostic — Use any provider or buy credits through Continue
Multi-surface platform — PR checks, CLI agent, and dual IDE support (VS Code + JetBrains)
CLI-first automation — Refine in TUI, deploy in Headless mode
Genuinely open source — Apache 2.0 core with 33K+ stars and ~3.3M VS Code installs
Active development — Repository actively maintained as of June 2026^[1]

Cautions

Strategic churn — The product has pivoted from IDE extension to "Continuous AI" platform to PR checks; positioning may shift again
IDE extension cadence — The VS Code extension's last marketplace update was March 2026, suggesting focus has moved to checks and CLI^[3]
Mixed extension reviews — 3.4/5 average rating on the VS Code Marketplace^[3]
Seed-stage funding — ~$5.1M raised is modest against competitors like Cursor and GitHub Copilot
Docs in flux — Documentation URLs have churned with the repositioning; some older links are dead

Pricing & Licensing

Tier	Price	Includes
Starter	Pay-as-you-go, $3/M tokens	Create and run agents, integrations (Slack, Sentry, Snyk), credits for frontier models
Team	$20/seat/mo	+ Private agent sharing, team management, usage controls, $10/seat credits
Company	Custom	+ SSO (SAML/OIDC), BYOK, commitments, invoicing, SLA

^[9]

The open-source CLI and IDE extensions remain free to run with your own API keys (Apache 2.0).^[1]

Licensing model: Open source tools with usage-based and per-seat cloud pricing. The former free "Solo" cloud tier has been replaced by pay-as-you-go token pricing as of June 2026.

Hidden costs: Token-metered checks running on every PR add up with PR volume. Budget accordingly.

Competitive Positioning

Direct Competitors

Competitor	Differentiation
GitHub Copilot	Copilot's code review is general-purpose; Continue's checks are repo-defined and deterministic in scope
CodeRabbit / Greptile	AI PR reviewers offering broad feedback; Continue enforces only the standards you write down
Cursor	Cursor is a full IDE (with Bugbot for PRs); Continue adds to your existing editor and CI
Cline	Both offer open-source coding agents; Continue adds source-controlled PR checks

When to Choose Continue Over Alternatives

Choose Continue when: You want repo-defined, enforceable AI quality gates on every PR, plus an open-source CLI agent
Choose Cursor when: You want a dedicated AI-native IDE experience
Choose GitHub Copilot when: You need deep Microsoft/GitHub integration
Choose Cline when: You want VS Code-only with richer MCP extensibility

Ideal Customer Profile

Best fit:

Teams that want codified, enforceable review standards on every PR
Teams building CI/CD pipelines with AI quality gates
Developers who work across IDE and terminal
Organizations needing both VS Code and JetBrains support
Users who value open source with commercial backing

Poor fit:

Individual developers wanting the simplest possible autocomplete
Teams satisfied with GitHub Copilot's existing integration
Users who primarily want a polished interactive IDE assistant — that is no longer the company's center of gravity
Teams with high PR volume and tight budgets (token-metered checks add up)

Viability Assessment

Factor	Assessment
Financial Health	Moderate — ~$5.1M seed from YC and Heavybit; no Series A announced as of June 2026^[6]
Market Position	Challenger — Strong OSS base, now competing in the crowded AI code review space
Innovation Pace	Rapid but redirected — Active repo, though IDE extension updates have slowed
Community/Ecosystem	Large — 33K+ stars, ~3.3M VS Code installs, community Discord^[1]^[3]
Long-term Outlook	Uncertain-positive — The checks niche is differentiated, but the pivot resets go-to-market

Continue has now pivoted twice — from IDE extension to "Continuous AI" agent platform to source-controlled PR checks. The current thesis is sharper: deterministic, repo-defined quality gates rather than another general AI reviewer. But seed-stage funding and a crowded code review market (CodeRabbit, Greptile, Copilot code review, Cursor Bugbot) make execution risk real.

Bottom Line

Continue has repositioned from a broad open-source coding assistant to a focused bet: source-controlled AI checks that enforce your team's written standards on every pull request, backed by an open-source CLI agent and IDE extensions that remain free to self-run. The differentiation is real — checks live in the repo and behave like CI, not like a chatty reviewer — but the company is now playing in the AI code review market rather than the coding assistant market.

Recommended for: Teams that want codified, enforceable AI quality gates in CI, especially those already invested in open-source tooling and multi-IDE environments.

Not recommended for: Developers seeking a polished interactive assistant or simple autocomplete — the company's energy is in PR checks, and the IDE extension's update cadence reflects that.

Outlook: The checks thesis is the sharpest positioning Continue has had, and the open-source CLI plus 33K-star community give it distribution. The risks are seed-stage funding, two pivots in two years, and well-funded incumbents (GitHub, Cursor, CodeRabbit) converging on PR review.

Research by Ry Walker Research • methodology

Sources