go_job
Job search, resume optimization, career research, and application tracking — a Go MCP server exposing 28 tools across LinkedIn, Greenhouse, Lever, YC, HN, Indeed, Хабр, RemoteOK, WeWorkRemotely, Twitter/X, Google Jobs, and more.
接入你的智能体
從發布者安裝。不是 PlugMCP 端點。
遠端 URL 還是本機指令? 兩種安裝方式的差異
來自發布者
Job search, resume optimization, career research, and application tracking — a Go MCP server exposing 28 tools across LinkedIn, Greenhouse, Lever, YC, HN, Indeed, Хабр, RemoteOK, WeWorkRemotely, Twitter/X, Google Jobs, and more.
Job, Remote & Freelance Search MCP server
| Tool | Description |
|---|---|
job_search | Unified search: LinkedIn, Greenhouse, Lever, YC, HN, Indeed, Хабр, RemoteOK, WeWorkRemotely, Remotive, Twitter/X, Google Jobs. platform= selects source; limit (default 15, max 50) + offset for pagination. |
job_match_score | Score job listings against a resume using Jaccard keyword overlap (0–100). |
opportunity_search | Cross-type opportunity search (jobs + freelance + bounty). |
opportunity_analyze | Deep analyze a single opportunity URL. |
opportunity_claim | Initiate a claim action on a matched opportunity. |
go_job/
├── main.go
├── internal/
│ ├── engine/
│ │ ├── config.go # Config struct + Init()
│ │ ├── bridge.go # Source bridge wiring
│ │ ├── bridge_jobs.go # Job-specific bridge helpers
│ │ ├── bridge_llm.go # LLM bridge
│ │ ├── cache.go # 2-tier cache: L1 in-memory + L2 Redis
│ │ ├── search.go # SearchSearXNG, DedupByDomain
│ │ ├── metrics.go # Prometheus counters/histograms
│ │ ├── pipeline.go # Fan-out pipeline helpers
│ │ ├── types_jobs.go # Input/output types
│ │ ├── prompt.go # LLM instructions (shared)
│ │ ├── prompt_jobs.go # Job-specific LLM prompts
│ │ └── jobs/ # Job source + career tool implementations
│ │ ├── linkedin.go # LinkedIn Guest API + JSON-LD + geo_id + pagination
│ │ ├── indeed.go # Indeed iOS GraphQL API + SearXNG fallback
│ │ ├── remotejobs.go # RemoteOK + WeWorkRemotely + Remotive APIs
│ │ ├── habr.go # Habr Карьера scraper
│ │ ├── hnjobs.go # HN Who is Hiring (Algolia)
│ │ ├── ycjobs.go # YC workatastartup.com
│ │ ├── ats.go # Greenhouse + Lever + Ashby ATSes
│ │ ├── match.go # Jaccard keyword scoring (job_match_score)
│ │ ├── resume.go # resume_analyze, cover_letter_generate, resume_tailor
│ │ ├── research.go # research tool (salary / company / person)
│ │ ├── tracker.go # Job application tracker (SQLite)
│ │ ├── profile.go # User profile persistence
│ │ └── ... # algora, twitter, linkedin, bounty, opportunity, etc.
│ │ └── sources/ # Pluggable source connectors
│ │ ├── freelancer.go # Freelancer.com REST API
│ │ ├── github.go # GitHub jobs / PRs
│ │ ├── hackernews.go # HN source
│ │ └── ...
│ ├── jobserver/
│ │ ├── register.go # Tool registrations (28 MCP tools)
│ │ └── tool_*.go # Per-tool handler files
│ └── hunt/ # Hunt store + notifications
│ ├── notify/
│ │ └── telegram.go # Telegram notifications via go-kit ProductSink
│ └── ...
└── deploy/
└── go_job.service # systemd unit (MCP :8891, metrics :9891)
remote_work_search and freelance_search and twitter_job_search are folded into job_search via platform=remoteok|weworkremotely|remotive|twitter. Use raw=true with platform=twitter to skip LLM processing.bogdanfinn/tls-client; pagination with offset; 42 geo locations.platform=google via SearXNG.platform=inspira (careers.un.org) / platform=undp / platform=un (fan-out both). Not included in platform=all.salary_research, company_research, person_research).subject=salary (role required), subject=company (company required), subject=person (name required).job_tracker_add, job_tracker_list, job_tracker_update.action=add (title+company required), action=list, action=update (id required).$UPLOADS_ROOT/go-job/tracker/tracker.db (SQLite, table jobs). Default path: $HOME/uploads/go-job/tracker/tracker.db. Override via UPLOADS_ROOT.$UPLOADS_ROOT/go-job/profile/profile.json. Default: $HOME/uploads/go-job/profile/profile.json.sync.Map), lost on restart.internal/hunt/notify/telegram.go using go-kit ProductSink (own bot, rate-limited fan-out).TELEGRAM_BOT_TOKEN + HUNT_NOTIFY_CHAT_ID env vars.# HTTP mode (default MCP port 8891, metrics port 9891)
MCP_PORT=8891 PROM_PORT=9891 LLM_API_KEY=... ./bin/go_job
# stdio mode
./bin/go_job --stdio
make build # → bin/go_job
make deploy # build + copy service + restart systemd unit
make restart # restart only
curl http://localhost:8891/health
# {"status":"ok","service":"go_job","version":"v1.21.0"}
#
# `version` is `git describe --tags` taken at build time — the release tag that
# release-please cut, never a number maintained by hand here. A build that
# cannot reach git reports "dev".
curl http://localhost:9891/metrics
Large MCP responses can exceed the MCP envelope limit (~25KB). go-job spills oversized payloads to a Postgres table oversize_responses (auto-migrated on startup).
When a tool response exceeds GO_JOB_OVERSIZE_THRESHOLD_BYTES (default 24576), the client receives a small envelope with oversize_id and a sample. Use the oversize tool to retrieve the full payload.
Requires DATABASE_URL. Gracefully falls back to direct response if unset.
工具
尚無工具清單。遠端伺服器會在首次健康檢查後取得工具清單。
常見問題
在本頁複製適用於 Cursor、Claude Desktop、Claude Code 或 VS Code 的 stdio 設定。它使用發布者的安裝方式,而不是 PlugMCP 代理伺服器。
這筆記錄未列出必要的環境變數。
Go Job 透過 stdio 在本機執行。
上次檢查時間:2026-11-15。 Job, Remote & Freelance Search MCP server