User repository on GitHub
Multi-agent AI news pipeline powered by GLM-5.3-Flash and Jev
- Forks
- 9
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
CommunityJev Agents & MCP
AI News Aggregator
Daily AI news pipeline that gathers RSS, arXiv, and social sources with staged agents and publishes a searchable web frontend.
- Stars
- 32
- Language
- Python
- License
- APACHE-2.0
- Published by
- Community
- Author
- flyryan
- Last verified
- Oct 2026
About this project
Daily AI/ML news briefings curated by specialized agents using adaptive thinking profiles. The publishing repository starts the hosted pipeline every morning at 3 AM ET, with the live site typically updated around 4 AM ET. A Python-based pipeline that collects AI/ML news from multiple sources, analyzes them using specialized agents with Claude's adaptive thinking, and serves a modern Svelte SPA frontend. 40+ curated RSS/Atom sources plus 7 arXiv category feeds.
Quoted from the project's README. Read the full README.
- Languages
- Python
Highlights
- Collects from dozens of RSS sources, multiple arXiv categories, and several social and research-blog platforms.
- Multi-phase pipeline covers gathering, batched analysis, continuity tracking, topics, summary, and enrichment.
- Produces JSON data, replay artifacts, RSS feeds, and a client-built search index for the Svelte frontend.
- Ships Docker, local, and web-only run options plus health, resume, and cookie-fetch helper scripts.
Watch out
Self-hosting needs provider keys, config files, and pipeline resources; the scheduled publishing workflow is guarded to the upstream repository. Copy reviewed against the repository README on 2026-10-04, not from running it.
Questions about this entry
Short answers drawn from the linked project or page.
RSS and arXiv gatherers plus social collectors feed batched analysis and synthesis, then topic detection, an executive summary, link enrichment, and a generated hero image before JSON, feed, and index output.
They are effort profiles for the reasoning model, from quick summarization through standard batch work up to deep ranking and cross-category synthesis, not fixed token budgets on current models.
Docker with a config file is the recommended path, local Python with a virtual environment works for development, and a lightweight web-only image serves just the frontend.
Forks can change the workflow guard, schedule, secrets, and base URL, and rerun from checkpoints or specific phases such as topic detection onward.
Forty-plus curated RSS and Atom sources, seven arXiv category feeds (cs.AI, cs.LG, cs.CL, cs.CV, cs.NE, cs.RO, stat.ML), and six social platforms: Twitter, Bluesky, Mastodon, Reddit, LessWrong and research blogs. GitHub Actions cron runs in UTC, so the workflow carries two UTC entries with a local-time guard and only the one mapping to 3 AM America/New_York proceeds; the live site is typically updated around 4 AM ET.
The analysis phase uses a MAP-REDUCE pattern: items batched 75 at a time, analyzed in parallel, then synthesized from STANDARD to DEEP. For Claude Opus 5 the profiles are adaptive rather than fixed token budgets, sent as thinking with type adaptive and display summarized and mapped to output_config.effort: QUICK at high effort, STANDARD at xhigh, DEEP and ULTRATHINK both at max. Manual budget_tokens is only used for older Claude models.
Apache-2.0, and it needs your own provider keys: PIPELINE_PROVIDERS_YAML holding the full ignored config/providers.yaml, ANTHROPIC_API_KEY, and optionally ANTHROPIC_API_BASE for an OpenAI-compatible proxy. Collection is optional per source: TWITTERAPI_IO_KEY for Twitter, SCRAPECREATORS_API_KEY for Reddit, which replaced the dead free Reddit .json endpoint and is required for Reddit data, and LESSWRONG_PROXY_URL, PIPELINE_PROXY_URL or a Mullvad account for egress.
More Jev resources like this
51GitHub stars
Local LinkedIn recruiting workspace on Browser Use's Jev Ultrafast: Jev screens titles, drives navigation, and assesses profile excerpts against your brief while a live browser stays visible; matches and evidence export as JSON.
Jev Agents & MCP#community#jevbest#catalog-import
Community
skeptrunedev/jev-recruiterPython
skeptrunedev/jev-recruiterPython33GitHub stars
Codex Skill plus stdlib-only Python CLI that recommends which installed skill to read next for a task, allowing no-skill or uncertain answers and never executing candidates.
Jev Agents & MCP#community#jevbest#catalog-import
Community
win4r/jev-skill-suggesterPython
win4r/jev-skill-suggesterPython72GitHub stars
MCP server over stdio that gives coding agents fourteen Jev judgment tools, turning evidence plus an enumerable question into auto, review, or escalate actions from calibrated probabilities.
Jev Agents & MCP#community#jevbest#catalog-import
Community
PyModel/jev-judge-mcpPython
PyModel/jev-judge-mcpPython21GitHub stars
Research runtime that pages a large logical tool and context space through a small bounded Jev decision surface with explicit approvals.
Jev Agents & MCP#community#jevbest#catalog-import
18GitHub stars
Python CLI and MCP server for keyless web search over DuckDuckGo or self-hosted SearXNG, with an optional local Laya sieve that filters, reranks and stance-checks evidence before it reaches an agent.
Jev Agents & MCP#community#jevbest#catalog-import
14GitHub stars
Tiered agent memory keeping compact summaries with source links over searchable raw history, where a router decides per query whether notes suffice or raw pages must be recovered.
Jev Agents & MCP#community#jevbest#catalog-import
Community
FreedomIntelligence/TiermemPython
FreedomIntelligence/TiermemPythonFrom the Jev community
Posts from builders shipping with Jev right now.