Run one real workflow first
Do not start by memorizing endpoints. Start by completing one real workflow: create an API key, call Context API, and expand only when the task needs deeper signals.
One recommended path, three actions
- Default 2-call mental model
- Start with Context API
- Expand only when needed
- Create an API key and establish integration identity.
- Call Context API and get the aggregated career context first.
- Only expand into reasoning, dynamic context, or webhooks when the task requires it.
Get your first successful call fast
Follow this order: identity first, context second, task layer third.
Step 1 — Create an API Key
Generate an API key in Settings so your agent or application has a stable integration identity.
Step 2 — Call Context API
Treat `/api/v1/context` as the default entrypoint. It gives you a strong aggregated result before you branch further.
Step 3 — Expand only when needed
Move into deeper graph, memory, or specialized endpoints only when the task proves you need them.
Treat expand as optional
The most common mistake is requesting the full object too early. Start with the smallest useful object, then deepen progressively.
Default
Send no `expand` parameter first and validate whether the default context is already sufficient.
Reasoning signals
Add `expand=careerGraph` only when you need more structured reasoning and explanation signals.
Query-specific memory
Move into `dynamicContext` only when the user's current question depends on recent memory and temporal context.
Learn the rules before the endpoints
Learn these principles first and the rest of the API becomes much easier to use well.
Start broad
Begin with the aggregated context layer instead of splitting into many narrow calls too early.
Stay small
Keep the default request compact and only add fields or expand values when the task proves you need them.
Branch by contract
Implement errors, auth, and webhook flows by formal contract instead of assumptions.
Your first request should stay small
Start with aggregated context. That keeps cost lower and gives downstream decision layers a more stable starting point.
curl "https://api.profileclaw.com/api/v1/context" -H "Authorization: Bearer $PROFILECLAW_API_KEY"curl "https://api.profileclaw.com/api/v1/context?expand=careerGraph&careerGraph=summary&resumeSkillsLimit=20" -H "Authorization: Bearer $PROFILECLAW_API_KEY"Once the first flow works, go deeper
Quickstart does not need to teach everything. It needs to establish the right mental model fast.