Cline Kanban hits provider-routing bug

// 143d agoPRODUCT UPDATE

Cline Kanban hits provider-routing bug

Official Cline docs describe Kanban as CLI-agnostic and compatible with Cline CLI, and Cline itself supports local models through Ollama, LM Studio, and OpenAI-compatible endpoints. That makes local LLM use a supported path in principle, so if Kanban is ignoring a llama.cpp-compatible base URL and trying to hit OpenAI directly, the most likely explanation is a bug or regression in how Kanban is wiring the agent/provider settings rather than an intentional cloud-only design.

// ANALYSIS

Hot take: this does not read like an “online models only” product; it reads like a config path that is being bypassed somewhere.

–Cline Kanban is documented as working with CLI-based agents, including Cline CLI, and Cline CLI already supports local/OpenAI-compatible providers.
–The official docs explicitly mention local models via Ollama and LM Studio, plus OpenAI-compatible endpoints.
–If Kanban is forcing OpenAI behavior despite a local base URL, that points to a bug in Kanban’s handoff to the agent runtime or provider selection.
–Recent community reports suggest similar OpenAI-compatible/local-model routing issues, which strengthens the bug hypothesis.
–Best interpretation: local LLM support should be possible, but this specific flow may still be broken and worth filing with logs.

// TAGS

clinekanbanlocal-firstllama.cppopenai-compatiblecoding-agentai-toolsai-coding

DISCOVERED

143d ago

2026-05-01

PUBLISHED

143d ago

2026-05-01

RELEVANCE

8/ 10

AUTHOR

PairOfRussels

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE19m ago

Synara 0.9.0 brings Computer Use to macOS

Synara released version 0.9.0 of its open-source, local-first coding workspace, introducing experimental Computer Use on macOS alongside window-targeted live previews. The update also adds local project and conversation imports from Claude Code and Codex, Claude prompt-cache inspection, and non-blocking question cards.

BENCHMARK2h ago

Jev 1.13 beats Cohere, Voyage in latency

Michael Chomsky shared latency benchmark results comparing TypeSafe AI's Jev 1.13 model against leading dedicated rerankers for search and retrieval pipelines. In the test, Jev 1.13 registered the fastest speeds with a 149 ms median and 205 ms p95 latency, beating Cohere v3.5 (176 ms median / 357 ms p95), Voyage 2.5 Lite (183 ms median / 232 ms p95), and Cohere 4 Fast (183 ms median / 824 ms p95).

MODEL2h ago

Tripo3D P2 lands on fal with decoupled controls

Tripo3D has launched its P2 model on fal, providing fast API-driven text-to-3D and image-to-3D textured mesh generation for creators and developers. The model introduces four texture-quality tiers, decoupled seeds for geometry and texture to facilitate fine-grained iterative design, and support for removing baked-in lighting via the v3.5 texture model for seamless integration into real-time 3D pipelines.