Hugging Face's logo

RMDWLLC
/
kaiju-coder-7-adapter

Text Generation
PEFT
Safetensors
English
kaiju-coder-7
lora
coding
local-ai
business
opencode
conversational
Model card Files
xet
 Community
kaiju-coder-7-adapter / PUBLIC_TESTING_QUICKSTART.md
restokes92's picture
restokes92
Upload Kaiju Coder 7 adapter release package
ee7de59 verified
preview code
|
raw
history blame contribute delete
6.96 kB

Kaiju Coder 7 Public Testing Quickstart

Kaiju Coder 7 is the public model name. The OpenAI-compatible model id is:

kaiju-coder-7

Use this guide for serious public testing. It avoids internal checkpoint names and keeps the current limitations clear.

Pick A Test Path

Path 1: OpenCode Against An Existing Endpoint

Use this if you already have Kaiju Coder 7 served at an OpenAI-compatible /v1 endpoint.

git clone https://huggingface.co/RMDWLLC/kaiju-coder-7-opencode
cd kaiju-coder-7-opencode
python3 scripts/install_kaiju_opencode_profile.py --base-url http://127.0.0.1:18181/v1

Then run OpenCode inside the project you want to edit:

opencode

The installer sets kaiju/kaiju-coder-7 as the OpenCode model and kaiju-coder-7 as the default agent. You can still select kaiju/kaiju-coder-7 manually from OpenCode's model picker if you switch away.

For a bounded smoke test:

mkdir -p /tmp/kaiju-public-smoke
opencode run --dir /tmp/kaiju-public-smoke \
  "Create hello.txt with exactly: Kaiju Coder 7 is ready"

Or run the packaged verifier, which checks the installer, live model endpoint, OpenCode binary, actual file creation, and wrong-directory behavior:

python3 scripts/run_kaiju_public_opencode_smoke.py

The helper installer adds:

  • the kaiju OpenAI-compatible provider
  • model: kaiju/kaiju-coder-7 and default_agent: kaiju-coder-7
  • the lean kaiju-coder-7 OpenCode agent
  • Kaiju as the default primary agent, so selecting Kaiju Coder 7 uses the hidden fast artifact path without requiring /kaiju
  • the kaiju-coder-7-run router command for fast websites, owner packs, and Desktop artifact folders
  • the kaiju_artifact OpenCode custom tool and /kaiju command for routing large artifact prompts through the fast local router
  • a scoped no-autocontinue plugin that prevents false completion loops after compaction or output limits

For a fast website or owner-pack artifact without waiting on raw OpenCode multi-file streaming, run:

kaiju-coder-7-run \
  --no-planner \
  --kind website \
  --out-dir "$HOME/Desktop/Kaiju-Coder-7-Test" \
  --prompt "Build a premium one-page website for Harborline Bookkeeping with pricing, FAQ, and a cleanup-call CTA."

OpenCode should use this same command internally for large website, business-pack, and Desktop-output requests after the helper is installed.

Inside OpenCode, /kaiju is optional for large generated artifacts. The command is prompt-backed, but it points the Kaiju agent at the kaiju_artifact custom tool instead of making the model hand-write every file.

Path 2: Full Local Weights

Use this if the full RMDWLLC/kaiju-coder-7 Hugging Face repo has been uploaded and you have suitable local GPU hardware.

hf download RMDWLLC/kaiju-coder-7 --local-dir ./kaiju-coder-7

Serve the downloaded folder with an OpenAI-compatible local server. Configure the server to expose:

model id: kaiju-coder-7
base URL: http://127.0.0.1:18084/v1
context: 16384

For the fastest OpenCode behavior, run the bundled fast proxy in a separate terminal and point OpenCode at the proxy:

KAIJU_OPENAI_BASE_URL=http://127.0.0.1:18084/v1 \
python3 scripts/kaiju_opencode_fast_proxy.py --host 127.0.0.1 --port 18181

Then install the OpenCode helper with:

git clone https://huggingface.co/RMDWLLC/kaiju-coder-7-opencode
cd kaiju-coder-7-opencode
python3 scripts/install_kaiju_opencode_profile.py --base-url http://127.0.0.1:18181/v1

Path 3: Runtime-Quantized Local Candidate

Use this only if you are comfortable with advanced serving setups. The current working quantized option is a runtime bitsandbytes recipe. A Q8_0 GGUF artifact has been converted, but it is still a candidate until runtime smoke passes.

git clone https://huggingface.co/RMDWLLC/kaiju-coder-7-quantized-runtime
cd kaiju-coder-7-quantized-runtime

Read README.md in that repo before serving. This path can reduce model memory at runtime, but it still depends on access to the full Kaiju Coder 7 weights.

Recommended Test Prompt

Run this from an empty project folder:

Build a launch-ready local service business website and operating pack. Include
index.html, a Stripe checkout safety plan, a CSV parser with tests, a simple CRM
schema, a weekly money report, and a safety/provenance note. Write the files,
not just advice.

Expected result:

  • files are written in the requested project folder
  • index.html is complete HTML
  • business docs start with Markdown H1 headings
  • code includes a test or smoke-check command where practical
  • no fake API keys, OAuth tokens, payment secrets, or private customer data

Current Recommended Defaults

  • Public model id: kaiju-coder-7
  • OpenCode context: 16384
  • Output cap for public testing: 2500
  • Fast OpenCode path: vLLM bitsandbytes runtime behind the Kaiju fast proxy
  • Current reliable product path: model plus deterministic business-owner harness/router plus verifier
  • Raw multi-file OpenCode generation: still too slow for broad paid claims; use kaiju-coder-7-run for fast public website and owner-pack tests while broader raw-model latency gates continue
  • Paid API: not public until launch preflight passes and the Stripe live-mode switch is deliberately completed

What Not To Claim Yet

Do not claim:

  • that raw model weights alone reliably build every business-owner artifact
  • that a paid hosted API is generally available
  • that persisted quantized weights exist
  • that 32k context is the current live default

Do claim:

  • Kaiju Coder 7 has a working local/OpenCode release candidate
  • the current tested OpenCode default is 16k context
  • the helper package includes a lean agent and compaction loop guard
  • the helper package includes the kaiju-coder-7-run router command for fast artifact generation
  • the fast proxy keeps OpenCode tool calls intact while forcing bounded, non-thinking generation
  • the paid API scaffold has tests and a launch preflight, but is not yet public
  • the packaged public smoke verifies a fresh OpenCode one-file write before public claims are refreshed
  • a GGUF Q8_0 candidate exists, but is not public quantized-weights release evidence until runtime smoke passes

Remaining Caveats Before Broader Claims

  • Hugging Face public release repos are uploaded and public under RMDWLLC.
  • The GGUF Q8_0 candidate still needs a runtime smoke before public quantized-weights upload.
  • Raw multi-file OpenCode generation is still not the public speed story; use the deterministic router/harness for websites and business-owner packs.
  • Public paid API launch has approval and preflight evidence, but real customer charging still needs a deliberate Stripe live-mode switch and controlled live payment verification.
  • Do not claim 32k context as the live default until it is freshly restarted and re-confirmed.