Price terms as published
Each line is the vendor’s own wording, split from the price capture checked on 2026-10-14. Nothing here is rounded, converted or re-worded — the full sentences stay in the review below. Lines that carry only a price continue the plan named on the line above, in the vendor’s own order.
| # | Plan / term | Checked |
|---|---|---|
| 1 | Free: signup credit ~$0.5 (tester-reported) + 14 zero-cost :free models | 2026-10-14 |
| 2 | paid = per-token gateway pricing | 2026-10-14 |
One-liner: Cline is an Apache-2.0 open-source coding agent for VS Code (github.com/cline/cline) that also runs a hosted model gateway at api.cline.bot. The gateway lists hundreds of models, marks a small subset :free, and doubles as Cline’s subscription billing pipe.
Verified facts (catalog and repo stats re-read 2026-10-14)
| Attribute | Value | Source |
|---|---|---|
| GitHub stars / forks | 70,284 / 7,655 | GitHub API, 2026-10-14 |
| License | Apache-2.0 | GitHub API, same day |
| Last push | 2026-10-14 (actively developed) | GitHub API, same day |
| Catalog size | 462 models returned by /api/v1/models (was 428 on 2026-09-08) |
direct API read, 2026-10-14 |
| Free subset | 14 models with :free suffix (was 16 on 2026-09-08) |
same response, counted programmatically |
| Base URL | https://api.cline.bot/api/v1 (OpenAI-compatible) |
official docs |
| Anthropic format | Not supported at the OpenAI endpoint (claim, untested) | secondary source; treat as unverified |
The :free lineup on 2026-10-14: google/gemma-4 (26B and 31B), the nvidia/nemotron-3 family (ultra/super/nano-omni/3.5-lightning/3.5-content-safety), poolside/laguna-s and laguna-xs 2.1, thinkingmachines/inkling + inkling-small, dots-studio/dots-3-note-preview, cohere/north-mini-code and liquid/lfm-2.5-2.6b — 14 IDs. Compare with the 16 we counted on 2026-09-08 and the two inclusionai/ling-3.0-flash variants are the entries that have left the list; the lineup itself moves on a weekly cadence, not a daily one.
What “free” does and doesn’t mean here
:freesuffix = zero billed cost at the API edge on the models carrying it that day. Flagship models circulating in Chinese social posts (GLM-5.3-flash, DeepSeek-V4, LongCat 2.0) are in the catalog but carry no:freemarker — the “free” claim for those applies to Cline’s own CLI-side promotions, not to API billing. Don’t copy the “20+ flagship models free” framing without checking the suffix.- Signup credit: new accounts reportedly receive ~$0.5 in credit, and community testing puts the free-model throughput around 150 requests/day. Neither number appears on a public official pricing page — both are tester-reported and will drift.
- No SLA on the free list. The
:freeset changes without notice (two entries in our same-day read had already rotated versus the day before). Build anything serious on a paid route.
The pattern this represents
Tool vendors are becoming second-tier free API sources. Cline (an IDE agent) and AMD (Radeon Cloud free API) both give away model access for different reasons: Cline’s gateway funnels VS Code users into its billing pipe; AMD’s funnels developers toward its GPUs. When a source says “free API,” the first question is which layer is actually free — the model, the account, or just the marketing.
Who it’s for
- Existing Cline users who want zero-config model access inside the extension
- Tinkerers who want an OpenAI-compatible endpoint with a rotating free subset and don’t care about stability
- Not a production gateway: no published rate-limit doc for
:freemodels, no uptime commitment
Sources
- https://api.cline.bot/api/v1/models (catalog +
:freecount, direct reads 2026-09-07, 2026-09-08 and 2026-10-14) - https://api.github.com/repos/cline/cline (stars/forks/license, same day)
- https://app.cline.bot/dashboard/account?tab=api-keys (key issuance, self-service)
- Community test notes 2026-09-06 (signup credit, ~150 req/day) — tester-reported, not official