ProfileClaw

ProfileClaw is a career profiling platform built by 职虾 that stores assessment signals, experience evidence, skill patterns, and preferences to provide a human-readable map and AI-readable structured context.

ProfileClaw offers four assessment tracks covering interests (Holland RIASEC), workplace behavioral styles (Big Five-derived), career drivers, and skills & talents, totaling 200+ questions and localized for the Chinese market.

Profile setup takes approximately 15 minutes as a one-time build, after which the profile persists for long-term reuse across AI agent interactions.

ProfileClaw integrates with AI agent runtimes including OpenClaw, Claude Code, Claude Work, and Nanobot, enabling agents to receive structured experience, preferences, limits, and signals from the user's profile.

Agent integration is provided through three paths: a skill.md URL that agents can read directly, an npm-based CLI (npx @profileclawai/skill install-skill), and a developer API key for programmatic access.

ProfileClaw encrypts profile data by default, keeps the encryption key with the user, and allows user-controlled, revocable access on a per-agent basis; every API call is authenticated and encrypted.

ProfileClaw's assessments are localized for the Chinese market and designed for real work contexts in China, built by the China-based entity 职虾.

ProfileClaw operates on a freemium model: the free tier includes one assessment with initial profile results, while a paid AI layer unlocks summaries, extraction, reasoning, and deeper analysis.

Users can upload a resume to ProfileClaw, and the paid AI layer converts it into structured profile signals readable by both humans and agents.

ProfileClaw targets individual users seeking reusable long-term context, frequent AI agent users needing a human-context layer, and teams and enterprises for talent operations.

ProfileClaw supports use cases spanning job search and transitions, freelance and client work, collaboration screening, content and personal brand development, advisory services, and enterprise recruiting (the latter explicitly marked as in progress).

ProfileClaw's core thesis is that a pre-built structured profile reduces AI token waste by compressing identity into stable fields, so agents spend tokens on execution instead of rediscovering who the user is through multi-turn chat.

ProfileClaw is a career profile platform by 职虾 that builds structured profiles from assessments, resume evidence, and user materials, designed for consumption by both humans and AI agents.

The product offers four assessment tracks: Holland RIASEC Interest Assessment, Workplace Behavioral Styles, Career Drivers, and Skills & Talents.

Profile setup is claimed to take approximately 15 minutes as a one-time build.

Profile_Protocol_v2 is referenced as the underlying protocol version.

Agent integration is provided through a skill.md endpoint at https://profileclaw.com/en/skill.md, with a recommended agent prompt that loads the profile.

A CLI fallback is available via npx @profileclawai/skill install-skill.

API key generation and a Context API are offered for developer access to the profile.

Encryption by default, user-controlled keys, and revocable access are claimed as security properties.

The product claims integration with OpenClaw, Claude Code/Claude Work, OpenCode, and Nanobot agent runtimes.

Assessments are localized for the China market with work-content adapted to Chinese professional contexts.

Resume upload is available with paid AI parsing to extract profile signals; extra materials (chats, screenshots, notes) can also be uploaded.

Enterprise recruiting features are described as in progress, with employer workflows using the same context layer under user permission.

User testimonials from three personas (recent graduate, career switcher, returnee) are presented with claimed outcomes including salary figures.

Documentation pages (Quickstart, API Overview, Docs, Pricing) are referenced in navigation but their actual content is not included in the frozen source pack.

The product is built by 职虾 (Zhixia) and the homepage is copyright 2026 ProfileClaw.

Documentation entry points span homepage CTAs, skill.md resource index, quickstart guides, API reference, developer guides sidebar, and OpenAPI at .well-known.

The product's value proposition is turning scattered career experience into structured, agent-consumable context.

The entry page was fetched and analyzed for machine-interface signals (title, headings, developer links, keyword probes).

Agent-native positioning with a concrete operational path: "The homepage provides a concrete agent-native workflow: a recommended agent prompt that reads a skill.md file and connects to ProfileClaw, plus CLI fallback commands for installing the skill.".

ProfileClaw publishes a public OpenAPI 3.1.0 specification at .well-known/openapi.json describing its REST API.

Beyond context, the API exposes profile, career graph, assessments, webhooks, career-prediction, similar-paths, and health endpoints.

The API returns standard rate limit headers including X-RateLimit-Remaining, X-RateLimit-Reset, and Retry-After.

The OpenAPI specification includes JSON response examples with data envelopes across multiple endpoints.

Error responses including 400, 401, 403, 404, 429, and 500 with x-error-codes are documented in the OpenAPI specification.

The API is versioned as v1, with version information declared in OpenAPI info and returned in response metadata.

Webhook subscription creation and event delivery with X-Webhook-Event headers are documented.

A POST /api/v1/career-prediction endpoint accepts a target role and returns transition feasibility with timeline estimates.

All API responses use structured JSON envelopes with data, metadata (requestId, timestamp, userId), and links fields.

The API exposes assessment data including RIASEC scores with dominant types and completion timestamps.

An endpoint /api/v1/similar-paths exists for career path comparisons.

ProfileClaw is positioned as a career context layer for AI agents and agent frameworks, providing structured profile and career graph data.

An OpenAPI 3.1.0 specification is published at /.well-known/openapi.json with API version declared as v1.

Webhook subscription lifecycle — create, list, and delete — is documented in the OpenAPI specification with event delivery headers.

Rate limiting is documented with X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset, and Retry-After response headers.

API responses follow a structured envelope pattern containing data, metadata (requestId, timestamp, userId), and links fields.

A career prediction endpoint POST /api/v1/career-prediction accepts a target role and returns a predicted fit analysis.

Assessment data includes RIASEC scores with dominant types, individual dimension scores, and completion timestamps.

Profile completeness is surfaced as a numeric score in API response metadata.

Error responses are documented with x-error-codes for HTTP 400, 401, 403, 404, 429, and 500 status codes in the OpenAPI specification.

ProfileClaw is positioned as a career context layer for AI agents and agent frameworks, delivering structured profiles, derived career graphs, compact agent-ready context, and webhook events via REST API.

Authentication to the API uses Bearer token or X-API-Key header; API keys are user-specific and generated through Settings → API Keys in the web interface.

The free tier is rate-limited to 100 requests per hour, with X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset headers returned on all responses.

Webhooks support six career-domain event types: profile.updated, assessment.completed, skill.acquired, goal.changed, insight.generated, and timeline.event.

The /api/v1/similar-paths endpoint computes career path similarity against a target role with a configurable minimum similarity threshold, returning ranked matches with duration and success metadata.

The API defines structured error responses with x-error-codes covering authentication, authorization, billing, and resource-not-found scenarios, each accompanied by rate-limit and retry headers.

The OpenAPI specification follows version 3.1.0 and includes typed schemas for profile, assessment, webhook configuration, and error response objects.

The API supports conditional requests via ETag and If-None-Match headers for caching.

A skill.md file provides agent integration guidance with priority-ordered endpoint recommendations.

The Context API endpoint GET /api/v1/context is designated as the primary, highest-priority entry point for agent integration.

ETag-based conditional requests with If-None-Match are documented for caching Context API responses.

A skill.md file provides agent implementation guidance covering API accessibility check, authentication flow, and data loading workflow.

The skill.md instructs agents to verify API accessibility, authenticate, and load available profile data immediately as the standard integration loop.

Authentication to the API uses Bearer token or X-API-Key header; API keys are user-specific and generated through Settings → API Keys in the web interface.

The free tier is rate-limited to 100 requests per hour, with X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset headers returned on all responses.

The Context API supports expand parameters including careerGraph and alerts, plus limit controls such as resumeSkillsLimit and alertsLimit for token-efficient agent consumption.

Agent skill definitions are published in both English and Chinese with localized trigger phrases and complete workflow instructions for each language.

API keys must never be exposed in logs and users are advised to delete and regenerate keys immediately if leakage is suspected.

Agent skill behavior rules instruct agents to reply in the user's language, prefer next actions over status-only updates, and avoid requiring the user to manually orchestrate the workflow.

The API supports Bearer token authentication via the Authorization header.

The core endpoint GET /api/v1/context provides aggregated career context as the recommended first call.

Documentation includes runnable curl request examples with query parameter and header variations.

Quickstart guides cover API key creation, Context API first call, and progressive workflow expansion.

API authentication supports Bearer token via the Authorization header.

The Context API supports expand parameters including careerGraph, dynamicContext, resumeSkillsLimit, and query for targeted data retrieval.

Quickstart documentation exists in both English and Chinese with parallel content.

curl request examples are provided across quickstart, skill.md, and OpenAPI x-codeSamples for multiple endpoints.

The canonical API host declared across all official documentation is api.profileclaw.com.

The quickstart workflow prescribes creating an API key first, then calling the Context API for an aggregated career snapshot before exploring other endpoints.

A quick-start / agent-skills documentation page is reachable at https://profileclaw.com/en/docs/quickstart.

SDKs are described as future-facing; the reference page mentions 'future SDKs' and 'SDK-facing artifacts' but no production SDK exists.

A health check endpoint exists at /api/v1/health.

The reference page advocates a single-contract-source philosophy with OpenAPI as canonical and docs as human-readable surfaces.

The API reference page articulates intent to publish the contract layer, reference, docs, and SDK-facing artifacts while keeping payments, admin, and model strategy private.

The reference page describes future SDKs as consumers of reusable response and resource shapes from the OpenAPI contract.

The API also supports X-API-Key header as an alternative authentication method.

The authentication guide recommends attaching credentials at the tool adapter layer rather than exposing secrets in prompt-visible headers.

API authentication also supports the X-API-Key header as an alternative credential method.

Documentation is available in both English and Simplified Chinese across guides, quickstarts, and the auth reference.

Documentation recommends a minimal two-call workflow starting with Context API, noting deeper scopes require internal/manual application.

Deeper API scopes beyond the Context API are described as currently restricted to internal testing or manual application rather than self-service.

Developer guides cover agents, integrations, authentication, caching, retries, and webhooks.

The documented default agent usage pattern is one Context API read plus one task-specific expansion call.

Developer guide pages cover authentication, caching, retries, webhooks, integrations, and assessments in a cross-linked documentation structure.

An OpenClaw integration guide is referenced in the developer documentation alongside general integration overviews.

llms.txt is published at the site root and readable.

sitemap.xml is reachable and lists site pages.

A documentation surface is reachable at https://profileclaw.com/en/developers.

An API documentation surface is reachable at https://profileclaw.com/en/docs/api.

A quick-start / agent-skills documentation page is reachable at https://profileclaw.com/en/assessment/skills-talents.

Agent skill definitions are published in both English and Chinese with localized trigger phrases and complete workflow instructions for each language.