Your career knowledge, accessible to any agent.
One store you own. Any AI agent — including your own — can use it to help you land the job. And recruiters’ agents can query it too, read-only and under your control.
One store. Two sides.
Your work knowledge lives in one place you own. MCP — the open standard for connecting AI agents to tools — lets agents read from it.
Your agent, working for you
Connect your own agent and put your evidence to work on your job hunt — fit checks, grounded cover letters, interview prep.
Their agent, accountable to you
Recruiters’ agents query your Twin read-only, scoped per share link, with every call on your activity log.
Your agent, working for you
Connect your Twin to your own agent over MCP. Your store does the remembering; your agent does the work.
Check your fit before you apply
Point your agent at a job description and your Twin scores you against it — strong, partial, or insufficient on each requirement, with the evidence behind every call.
aigis_extract_jd_requirementsaigis_score_against_jdCover letters grounded in what you’ve done
No more pasting your CV into a chatbot and getting invented achievements back. Your agent pulls your real narrative themes and the evidence behind them, then drafts a letter you can stand behind.
aigis_get_narrative_anchorsaigis_query_knowledgeaigis_get_evidence_forWalk into interviews prepared
Ask your agent for the questions a given role is likely to probe, and the strongest evidence from your own history to answer each one.
aigis_prepare_for_roleaigis_get_readinessTailor every application without starting over
Your store holds everything; your agent surfaces the slice that matters for this role — the right projects, metrics, and phrasing.
aigis_query_knowledgeaigis_get_profileSee your gaps honestly
The same categorical scoring recruiters see works for you: where you’re “insufficient” for a role is exactly what to shore up — or a signal to aim elsewhere.
aigis_score_against_jdA store that gets better as you work
Drop in a project recap or a win as a journal entry — in the app, or proposed by your agent for your approval. It joins your knowledge base, and every future application benefits.
aigis_submit_journalaigis_submit_artifactThe workflow it replaces
Pasting your CV and a job description into a chatbot gives you generic letters and confidently invented achievements. Aigis grounds the same task in your real evidence and refuses to fabricate — the same honesty contract the recruiter side runs on.
“Draft a cover letter for this product role, using my real experience.”
Your agent calls your Twin — pulls your narrative anchors, queries the matching evidence, checks fit against the JD — and returns a letter built only from things you’ve actually done. Every claim traceable. Edit and send.
Why MCP, specifically
One source of truth, not seven copies
Your experience usually lives scattered — CV_v7_final.docx, LinkedIn, a portfolio, your memory. Aigis holds the canonical version; MCP lets any agent read from it.
Not locked in our UI
Because it’s MCP, your knowledge works in Claude, ChatGPT, or your own agent. Aigis is the store; you pick the tool.
Grounded every time
Your agent queries the source instead of re-summarising a pasted CV, so outputs stay consistent and evidence-backed — not freshly invented each session.
It’s yours, and it travels
You own the store, you revoke access in one click, and it follows you from one job search to the next.
Built for AI agents, not just humans
Aigis speaks MCP. Plug a Twin into Claude, ChatGPT, or your own agent. Seventeen tools, each scope-limited per share link, read-only by design for recruiters — and every call lands in your activity log.
Recruiters get read-only access. Write operations stay with you.
fact_check
Fact-check
Discovery tools plus knowledge query and evidence-for.
jd_analysis
JD analysis
Everything in fact-check, plus JD extraction and scoring.
full_read
Full read
Every read tool — the default for new share links.
The seventeen tools
Ask & converse
- aigis_ask_candidateAsk the Twin a single question; RAG-grounded answer with citations.
- aigis_chat_with_twinMulti-turn conversation with the Twin.
Verify & assess
- aigis_get_evidence_forRecruiter-grade fact-check: supporting evidence for a claim, or 'no evidence'.
- aigis_extract_jd_requirementsStage 1: parse a job description into structured requirements.
- aigis_score_against_jdStage 2: classify each requirement strong / partial / insufficient, with cited quotes.
- aigis_prepare_for_roleCandidate-side JD prep: likely questions and strongest evidence per requirement.
- aigis_get_readinessHow well-trained the Twin is — readiness level and source counts.
Profile & knowledge
- aigis_get_profileStructured profile: role, skills, experience, projects, certifications.
- aigis_get_knowledge_summaryOverview of the knowledge base: doc counts, interviews, journals, freshness.
- aigis_query_knowledgeRetrieve top-k raw knowledge chunks with scores and sources.
- aigis_get_certificationsCertifications with issuer, dates, credential URLs, verification status.
- aigis_get_availabilityJob preferences and availability (role, location, salary, active status).
- aigis_get_narrative_anchorsRecurring themes and stories — cover-letter feedstock with cited evidence.
- aigis_get_twin_capabilitiesWhat this Twin supports: tools, knowledge sources, protocol version.
Contribute
- aigis_submit_artifactSubmit a knowledge artifact; queued for the candidate's review.
- aigis_submit_journalSubmit a work-journal entry; queued for review before indexing.
Identity
- aigis_whoamiThe calling token's identity, scopes, agent type, and expiry.
Your work history shouldn’t live in seven documents.
Build one store you own — and bring whichever agent you like.