feat: add Zima as privacy-preserving OpenAI-compatible provider

[abhigyan-zima]

Hey, been using Zima for a while and kept wiring it into config manually, figured I'd just upstream it. Zima is OpenAI-compatible so the diff is small (~10 lines)... Zima doesn't store any of my data or LLM usage — they have special hardware and infrastructure so Zima as a provider can't see the prompts, inputs, or outputs...

---

Summary by cubic

Add zima as an OpenAI-compatible provider with gpt-4o and gpt-4o-mini, served via Zima’s privacy-preserving infrastructure. Enables chat (and autocomplete for gpt-4o-mini) with 128k context.

^{Written for commit 412822d. Summary will update on new commits. Review in cubic}

[github-actions]

Thank you for your submission, we really appreciate it. Like many open-source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution. You can sign the CLA by just posting a Pull Request Comment same as the below format.

---

I have read the CLA Document and I hereby sign the CLA

---

_{You can retrigger this bot by commenting recheck in this Pull Request. Posted by the CLA Assistant Lite bot.}

[cubic-dev-ai]

1 issue found across 1 file

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="packages/llm-info/src/providers/zima.ts">

<violation number="1" location="packages/llm-info/src/providers/zima.ts:11">
P1: Model regex `/gpt-4o/i` is a non-anchored substring match placed before `/gpt-4o-mini/i`, so `findLlmInfo('gpt-4o-mini')` matches the `gpt-4o` entry first and returns wrong metadata (missing 'autocomplete' recommendation, wrong display name). Anchor the regex to avoid collision: `/^gpt-4o$/i`.</violation>
</file>

<sub>Reply with feedback, questions, or to request a fix.

Fix all with cubic | Re-trigger cubic</sub>

[cubic-dev-ai]

P1: Model regex /gpt-4o/i is a non-anchored substring match placed before /gpt-4o-mini/i, so findLlmInfo('gpt-4o-mini') matches the gpt-4o entry first and returns wrong metadata (missing 'autocomplete' recommendation, wrong display name). Anchor the regex to avoid collision: /^gpt-4o$/i.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At packages/llm-info/src/providers/zima.ts, line 11:

<comment>Model regex `/gpt-4o/i` is a non-anchored substring match placed before `/gpt-4o-mini/i`, so `findLlmInfo('gpt-4o-mini')` matches the `gpt-4o` entry first and returns wrong metadata (missing 'autocomplete' recommendation, wrong display name). Anchor the regex to avoid collision: `/^gpt-4o$/i`.</comment>

<file context>
@@ -0,0 +1,26 @@
+      contextLength: 128000,
+      description:
+        "OpenAI's GPT-4o served through Zima's privacy-preserving infrastructure. Zima's hardware ensures prompts and outputs are never visible to the provider.",
+      regex: /gpt-4o/i,
+      recommendedFor: ["chat"],
+    },
</file context>

Suggested change

	regex: /gpt-4o/i,
	+ regex: /^gpt-4o$/i,