KnightLeap's in-app AI features run on your own provider key, the same
bring-your-own-AI idea as the MCP server. Your account, your model, your cost. You add a key once
under Settings → AI, and it's stored encrypted and never shown again.
Pick one provider below.
What your key powers. Once a key is set, you unlock the AI features throughout
KnightLeap — the daily brief on Today, story breakdown
(turn a story into tasks & acceptance criteria), backlog grooming (AI suggests
and scores stories toward your OKRs), sprint composition, and
objective check-in drafts. Every one is a suggestion you review before it's saved.
Settings → AI — pick your provider, paste your key, and save. Only the last four characters are ever shown back to you.
Anthropic (Claude)
Go to console.anthropic.com
and sign in (or create an account). This is the developer console, separate from your
Claude.ai chat subscription.
Open Plans & Billing and add a payment method or buy credits. API usage is
pay-as-you-go and billed by Anthropic, not by KnightLeap. A check-in draft costs a fraction of a cent.
Go to API keys → Create Key, give it a name
(e.g. “KnightLeap”), and copy it. It starts with
sk-ant- and is shown only once.
In KnightLeap, open Settings → AI, choose
Anthropic (Claude) as the provider, paste the key, and save.
Leave Model blank to use our sensible default
(claude-haiku-4-5-20251001), or pin a
specific Claude model id if you prefer.
OpenAI (ChatGPT)
Go to platform.openai.com
and sign in. This is the API platform, separate from a ChatGPT Plus subscription —
they're billed independently.
Open Settings → Billing and add a payment method or credits. API calls are
pay-as-you-go, billed by OpenAI.
Go to API keys
→ Create new secret key, name it, and copy it. It starts with
sk- and is shown only once.
In KnightLeap, open Settings → AI, choose OpenAI as the
provider, paste the key, and save.
Leave Model blank to use our default
(gpt-4o-mini), or pin another
chat-completions model id.
Google (Gemini)
Go to aistudio.google.com/apikey
(Google AI Studio) and sign in with your Google account. This is Google's developer surface for the
Gemini API, separate from the Gemini chat app.
Click Create API key. A new key works on Google's free tier; for higher rate limits,
enable billing on the linked Google Cloud project. API usage is billed by Google, not by KnightLeap.
Copy the key — it starts with AIza and is shown only once.
In KnightLeap, open Settings → AI, choose Google (Gemini) as the
provider, paste the key, and save.
Leave Model blank to use our default
(gemini-2.0-flash), or pin another
Gemini model id (e.g. gemini-2.5-flash).
Is it safe to share your key?
A fair question for any third-party app — here's the honest picture. Your key is stored
encrypted at rest and never shown again (only the last four characters, with a
last used timestamp so you can spot anything you didn't trigger).
You can revoke it any time from Settings → AI, and anything using
it stops immediately. Practically, it's the same trust you already place in any server, CI pipeline, or
hosting platform that holds a key in an environment variable — with better hygiene than a plaintext
.env file.
For extra peace of mind — good practice with any third-party app — lock the key down:
Use a dedicated key for KnightLeap, separate from the ones you use elsewhere, so you can revoke it in isolation.
Set a monthly spend cap and tight rate limits on that key's workspace/project, so usage is bounded no matter what.
Watch the usage in your provider's console — since the key is dedicated, anything you didn't initiate stands out immediately.
Rather not share a key at all? The AI features are entirely optional — the rest of KnightLeap works
without one.
Good to know
Stored encrypted. Your key is encrypted at rest and never rendered back —
we show only the last four characters so you know which key is set. Treat it like a password.
You're in control of cost. Usage is billed by your provider, on your account.
Remove the key any time from Settings → AI to switch it off.
Draft, never act. AI output is always a suggestion you edit before saving.
For example, a check-in draft proposes a confidence score — but you set it.