DevConsensus
Coding tools / Cline
IDE extension · Open source · HN opinions Nov 2025 – Nov 2026

Cline: what developers actually say

We read every Hacker News comment from Nov 2025 – Nov 2026 that gave a real opinion on Cline. 30 of the 35 opinions come from people who say they use it themselves.

73%approval, all opinions
–approval, last 90 days
#7of 22 tools ranked
28different developers

Cline has 73% approval across 35 opinions from 28 authors, and 30 of the opinions are firsthand. Cost and usage limits get the most consistent praise, at 6 positive to 1 negative and 4 positive to 0 negative. Reliability is the weak spot, with 3 positive and 5 negative comments about slow runs, plan mode still making edits, and loops on complex diffs. Monthly approval has swung widely: 100 in July, 75 in August, 100 in September, then 50 in October on 3 opinions.

Pick it if

  • Want to switch model providers mid-task
  • Care about cost and usage limits
  • Use VS Code and want an open-source extension

Look elsewhere if

  • Need reliable complex refactors and diff application
  • Expect plan mode to block all edits
  • Need fast responses or efficient token use
Visit Cline ↗

Strengths and weaknesses

Each row counts opinions that mentioned the topic. Bars to the right of the line are praise, bars to the left are complaints.

PraiseComplaints
Reliability
5
3
Cost & pricing
1
6
Code quality
1
3
UX
1
3
Usage limits
4
Customization
3

How opinion has moved

050100Dec '25Mar '26May '26Oct '26
Monthly approval (line) and number of opinions (bars). Months with fewer than 3 opinions are left out.

What developers say

Praise

Recommends Cline's pass subscription over Zai and OpenCode.HN comment · Oct 17, 2026
Switched from OpenCode to Cline and prefers its interface.HN comment · Jul 12, 2026
Uses Cline with VS Code and Mimo for big refactors and says it rarely makes mistakes.HN comment · May 26, 2026
Says Cline works fine out of the box for most cases, especially with a provider that has good KV caching.HN comment · Apr 28, 2026
Says Cline is the only way they can do serious development with Gemini models.HN comment · Apr 5, 2026
Switched to Cline with OpenRouter, using any model for very little money; no account needed.HN comment · Apr 2, 2026
Found Cline pretty good; it keeps their setup and lets them switch providers mid-task.HN comment · Mar 7, 2026
Says Cline one-shots an app almost every time they use it.HN comment · Mar 5, 2026

Complaints

Says Cline's plan mode still made changes, which motivated building their own harness.HN comment · Oct 19, 2026
Says Claude-backed Cline made many errors on a Spring Boot 3-to-4 migration and could not fix its own mistakes.HN comment · Jun 21, 2026
Built an MCP server to address LLM mistakes they saw in Cline and Claude Code.HN comment · May 24, 2026
Tried Cline but couldn't get it working well, partly because it expected OpenAI's output format.HN comment · May 11, 2026
Says Cline felt frustratingly slow, which led them to dig into its internals and rewrite parts of it.HN comment · Apr 27, 2026
Says Cline didn't work for their CLI-centric, approve-every-edit workflow.HN comment · Mar 31, 2026
Says Cline can't guide models well enough, so they loop when applying complex diffs.HN comment · Dec 9, 2025
Says Cline has major token-utilization problems.HN comment · Apr 22, 2026

Summaries are paraphrased. Each links to the original comment.

Cline compared

Best Cline alternatives, according to developers who switched →

FAQ

Is Cline worth it?

Cline has 73% approval across 35 opinions from 28 authors, with 30 firsthand reports. Users praise cost, usage limits, and provider switching. Reliability is the main concern, at 3 positive and 5 negative comments.

Is Cline better than OpenCode?

Two people preferred Cline to OpenCode. Two others preferred Roo Code instead. The counts are small, so they do not settle the comparison.

Is Cline good for large refactors?

One user reports big refactors with Cline, VS Code, and Mimo, and says it rarely makes mistakes. Another reports many errors on a Spring Boot 3-to-4 migration with Claude and says Cline could not fix its own mistakes. The data shows mixed results that depend on the model and task.