Job Hunter

Guides/API keys

How to get an API key

A directory for bring-your-own AI: why Job Hunter never stores your key, then setup steps, resource links, and pricing snapshots for every provider the app supports.

Why the key only lives in this tab

Job Hunter is a local-first app. When you paste an API key, it is held in this browser tab’s memory for the current session so the app can call your chosen provider from your machine. It is not uploaded to our servers, written to disk, or saved in IndexedDB with your profile and jobs.

Closing or refreshing the tab clears the key. That is intentional: a shared laptop, browser sync backup, or exported database dump should not leave a reusable credential behind. The tradeoff is that you re-enter the key when you start a new session—safer than persisting secrets by default.

You still send prompts to the provider you pick (OpenAI, Anthropic, Google, Groq, or a local Ollama endpoint). Review that provider’s privacy policy and set billing limits before heavy use.

OpenAI

  1. Create or sign in to an OpenAI Platform account (API billing is separate from ChatGPT Plus).
  2. Open API keys, create a secret key, and copy it once—you will not see the full value again.
  3. Add a payment method and optional usage limits so a runaway loop cannot surprise-bill you.
  4. In Job Hunter, choose an OpenAI model, paste the key, and keep the default base URL unless you use a proxy.

Resource directory

Pricing snapshot

Standard API rates in USD per 1M tokens. Convenience snapshot only—prices change. Confirm on the official pricing page.

ModelInputCached inputOutput
GPT-4o minigpt-4o-mini$0.15$0.075$0.60
GPT-4ogpt-4o$2.50$1.25$10.00
GPT-4.1gpt-4.1$2.00$0.50$8.00
GPT-4.1 minigpt-4.1-mini$0.40$0.10$1.60
o4-minio4-mini$1.10$0.275$4.40

Prefer mini models for resume/cover-letter drafts when quality is good enough. Set a monthly budget in billing either way.

Anthropic (Claude)

  1. Sign in to the Claude Console (API access is separate from Claude.ai Pro / Max).
  2. Open API keys, create a key starting with sk-ant-, and store it somewhere safe—you only see it once.
  3. Add billing and review spend controls. Haiku is usually enough for drafting; Opus costs more.
  4. In Job Hunter, pick a Claude model and paste the key. Keep https://api.anthropic.com/v1 unless you use a proxy.

Resource directory

Pricing snapshot

Standard API rates in USD per 1M tokens. Convenience snapshot only—prices change. Confirm on the official pricing page.

ModelInputCached inputOutputNotes
Claude Haiku 4.5claude-haiku-4-5$1.00$0.10$5.00—
Claude Sonnet 4.5claude-sonnet-4-5$3.00$0.30$15.00—
Claude Opus 4.5claude-opus-4-5$5.00$0.50$25.00—
Claude Sonnet 5 (intro)claude-sonnet-5$2.00$0.20$10.00Intro pricing through Aug 31, 2026; then $3 / $15

Cache-hit column is prompt-cache refresh pricing. Batch API is typically ~50% off standard rates. Job Hunter lists the 4.5 family; Sonnet 5 is shown as a newer Console option.

Google (Gemini / AI Studio)

  1. Open Google AI Studio → API keys and sign in with a Google account.
  2. Create a key (new keys are auth keys by default). Copy it into Job Hunter for this session only—do not commit it to git.
  3. Start on the free tier if you are experimenting. When you need higher limits, link Cloud billing and set budget alerts.
  4. Choose Gemini 2.5 Flash or Pro in Job Hunter. Restrict leaked or dormant keys in AI Studio; Google is phasing out unrestricted standard keys (full cutover planned for September 2026).

Resource directory

Pricing snapshot (paid tier)

Standard API rates in USD per 1M tokens. Convenience snapshot only—prices change. Confirm on the official pricing page.

ModelInputCached inputOutputNotes
Gemini 2.5 Flashgemini-2.5-flash$0.30$0.03$2.50Free tier available (rate-limited)
Gemini 2.5 Pro (≤200k)gemini-2.5-pro$1.25—$10.00Prompts up to 200k tokens
Gemini 2.5 Pro (>200k)gemini-2.5-pro$2.50—$15.00Long-context prompts over 200k tokens

Free-tier prompts may be used to improve Google products; paid tier generally is not. Your live RPM/TPM/RPD limits live in AI Studio—not in blog posts.

Groq

  1. Create an account at the GroqCloud console.
  2. Open API Keys, create a gsk-… key, and paste it into Job Hunter. Free tier needs no card.
  3. Pick llama-3.3-70b-versatile (or another Groq model) and keep base URL https://api.groq.com/openai/v1.
  4. Watch org-wide rate limits—extra keys do not multiply free quota. Upgrade billing when you outgrow them.

Resource directory

Pricing snapshot

Standard API rates in USD per 1M tokens. Convenience snapshot only—prices change. Confirm on the official pricing page.

ModelInputOutputNotes
Llama 3.1 8B Instantllama-3.1-8b-instant$0.05$0.08~840 tok/s — cheapest Groq text option
Llama 4 Scoutmeta-llama/llama-4-scout-17b-16e-instruct$0.11$0.34~594 tok/s
Llama 3.3 70B Versatilellama-3.3-70b-versatile$0.59$0.79Model listed in Job Hunter (~394 tok/s)
GPT-OSS 20Bopenai/gpt-oss-20b$0.075$0.30~1,000 tok/s

Groq’s pitch is speed (hundreds of tokens/sec) at low $/1M. Batch API is about half of on-demand. Confirm live rates on groq.com/pricing.

Ollama (local, no cloud key)

Prefer that resume text never leave your machine? Run a model locally. Job Hunter talks to Ollama’s OpenAI-compatible endpoint; the “API key” field can be any placeholder (for example ollama) because local auth is unused.

  1. Download and install Ollama, then start the app so the local server is running.
  2. Pull a model in a terminal, for example ollama pull llama3.2 or ollama pull mistral.
  3. In Job Hunter, choose Llama 3.2 or Mistral (Ollama). Keep base URL http://localhost:11434/v1 (note the /v1 for OpenAI compatibility).
  4. Enter any non-empty key value and save. Generation stays on your machine—your GPU/CPU and disk space are the real cost.

Resource directory

Cost snapshot

ItemCostNotes
Ollama software$0Free to download and run
Model weights$0Download once; uses disk (often several GB each)
InferenceYour hardwareElectricity + time; no per-token cloud bill

Quality and speed depend on your machine. Smaller models (8B-class) are friendlier on laptops; larger ones need more RAM/VRAM.

Quick info

Copy the details forms keep asking for — portfolio, GitHub, LinkedIn, and more.

Loading profile…

API key

Your API key, models, and endpoint are kept only in this browser tab’s memory. They are never written to IndexedDB. Closing or refreshing the page clears session overrides. For local development you can seed defaults with NEXT_PUBLIC_AI_API_KEY, NEXT_PUBLIC_AI_MODEL, and NEXT_PUBLIC_AI_QUALITY_MODEL in .env.local.

More info — how we keep keys temporary, where to get one, and pricing.

Cheaper model for job parsing, fit scoring, and profile import.

Higher-quality model used only when generating a tailored resume and cover letter.

Fills in from the everyday model. Override for a proxy or custom local endpoint (Ollama defaults to http://localhost:11434/v1).