Stelladocs
Examples & Recipes

Examples & recipes

Ready-to-paste configuration files organized by use case — cost-tier pipeline profiles, provider setups from coding-plan subscriptions to air-gapped local servers, and team-shared config.

Every page in this section is built around a complete, copy-pasteable configuration file — pick the page that matches your situation, paste, verify, adjust.

Setups by use case

Your situationRecipe
One API key, want the best default setupSingle provider, single key
Z.ai GLM coding-plan subscriberZ.ai coding plan
One OpenRouter key for every model familyOne gateway key
Ollama / vLLM / LM Studio, or no network at allLocal & air-gapped
Models must flow through GCP / AWSEnterprise cloud
Sharing .stella/ config across a teamTeam-shared settings

Cost-tier profiles

The inference pipeline makes several model calls per run, and each one is independently configurable. Three ready-to-paste tiers — and below, the cost model to reason about your own:

ProfileWorkerJudgeTriageTypical run*
Maximum qualityclaude-fable-5gpt-5.5deepseek-chat~$0.40 – $2.50
Balancedglm-5.2claude-fable-5deepseek-chat~$0.05 – $0.50
Dirt cheapdeepseek-chatglm-5.2deepseek-chat~$0.01 – $0.10

*Estimates for a medium task (a focused multi-file change, one revision) at the catalog's list prices, with prompt caching working normally. A lookup costs far less; a long refactor costs more. Your actual numbers live in stella stats and the Observatory. The three profiles assume keys for two or three providers — with one key, start from single-key instead.

What a pipeline run is made of

Cost intuition starts with knowing who calls the model, and how much:

StageCallsTypical sizeCost driver
Triage1~1–2k in, ~1 token outNegligible on any model — but why pay flagship rates for one token? Route it to the cheapest model you have.
Plan1 (+1 repair, rarely)~5–15k in, ~1k outSmall. Rides the worker's settings.
Witness~1 engine turn~10–30k in, ~1–2k outOnly when no --test-command was given. Test runs themselves are local and free.
Execute5–40+ steps~100–500k in (mostly cached), ~5–30k outThe dominant cost. The worker model's price — especially its output price — is most of your bill.
Verify0The deterministic ladder (flip oracle, diff budget) is free.
Judge0–1 per verification~3–10k in, ~0.5–1k outOnly when deterministic evidence is inconclusive. Cents, even on a flagship.
Revise0–2 extra execute roundslike Execute, smallerEach revision re-runs part of the execute cost.

Three consequences worth internalizing:

  1. The worker's output price dominates. Execute emits the most tokens by far. Moving the worker from $15/Mtok output to $2.20 (Fable 5 → GLM-5.2) changes your bill roughly 5–7×; moving the judge barely changes it at all.
  2. A flagship judge is cheap insurance. One judge call is a few thousand input tokens and under a thousand out — around 1–2¢ on gpt-5.5, 2–3¢ on claude-fable-5. There is almost never a good reason to skimp on the judge; there is every reason to make it a different family than the worker. Arm --test-command and many runs skip the judge entirely.
  3. Prompt caching is doing heavy lifting. Stella keeps the system prefix byte-stable specifically so most execute-step input bills at the cached rate (4–10× cheaper, depending on the model). That is why real runs cost far less than naive input tokens × list price math suggests. See Context Engine.

Multipliers

  • Goal mode repeats pipeline rounds until the judge signs off — up to 8 rounds. Budget for a multiple of the single-run estimate, and cap it: --budget 5.00.
  • Fleets run one pipeline per task. --budget is divided across the concurrency width and enforced fleet-wide, so the cap you set is the cap you spend.
  • Best-of-N candidates multiply execute cost by N — strictly opt-in.

The safety net

Whatever profile you run, --budget <usd> is a hard cap enforced between steps — a run aborts cleanly rather than overshooting:

stella --budget 2.00 run "migrate the config loader to the new schema"

And everything you actually spent is queryable afterwards — per model, per run, per resolved task — with stella stats and stella observe. Let the observed $/resolved number, not the sticker price, pick your lineup: see the model guide.

Stella is totally free and open source, dual-licensed MIT or Apache 2.0— your choice: MIT's three-paragraph simplicity, or Apache 2.0's explicit patent grant that enterprises prefer. Use it, fork it, embed it, ship it — no account, no phone-home, no strings. What that means →

On this page