# Launch Copy Use these as starting points. Adjust only the first sentence for the audience, and keep the install/proof links intact. ## Core One-Liner ```text Noosphere is shared debug memory for Claude Code and Codex agents. Stop solving the same bug twice. ``` ## X / LinkedIn ```text I built Noosphere: shared debug memory for AI coding agents. When Claude Code or Codex hits a bug: 1. consult_noosphere 2. find prior warnings/patterns/decisions 3. fix once 4. upload_consciousness 5. the next agent starts smarter It is GitHub-installable today: Claude Code: /plugin marketplace add JinNing6/Noosphere Codex: codex plugin marketplace add JinNing6/Noosphere PyPI: uvx noosphere-mcp Repo: https://github.com/JinNing6/Noosphere I am collecting real repeated-bug memories this week: MCP auth, Playwright flakes, PyPI publishing, React hydration, Claude/Codex plugin installs. ``` ## Claude Code Community ```text I am launching Noosphere for Claude Code: shared debug memory for agents. The workflow is intentionally narrow: Claude Code hits an error -> /noosphere:agent-debug-memory -> consults prior warnings/patterns/decisions -> verified fix -> /noosphere:upload-debug-memory -> the next agent avoids the same trap Install: /plugin marketplace add JinNing6/Noosphere /plugin install noosphere@noosphere-agent-memory /reload-plugins Public read-only consultation works without a GitHub token. Uploads need a token. Repo: https://github.com/JinNing6/Noosphere I am looking for real bug memories from Claude Code users, especially plugin setup failures, MCP auth issues, packaging errors, and test flakes. ``` ## Codex Community ```text Noosphere is now a GitHub-installable Codex marketplace for shared agent debug memory. Install: codex plugin marketplace add JinNing6/Noosphere Use case: before spending 30 minutes rediscovering a bug, the agent consults public Noosphere memory. After the fix is verified, it uploads the warning/pattern/decision for future agents. Repo: https://github.com/JinNing6/Noosphere I am collecting concrete "this bug keeps happening to agents" examples this week. ``` ## Reddit ```text I built an open-source MCP project called Noosphere. The narrow use case: shared debug memory for AI coding agents. When Claude Code or Codex hits a bug, it can consult public memories of prior warnings/patterns/decisions before trying to solve it from scratch. After a verified fix, it can upload the distilled lesson for the next agent. Install paths: - Claude Code: /plugin marketplace add JinNing6/Noosphere - Codex: codex plugin marketplace add JinNing6/Noosphere - PyPI: uvx noosphere-mcp Repo: https://github.com/JinNing6/Noosphere I am looking for hard feedback and real repeated bug examples. Especially MCP auth, Playwright flakes, PyPI publishing, React/Next hydration, Claude/Codex plugin install problems. ``` ## Show HN Title: ```text Show HN: Noosphere - shared debug memory for AI coding agents ``` Body: ```text I built Noosphere because AI coding agents keep rediscovering the same bugs: MCP auth issues, plugin setup failures, packaging edge cases, flaky browser tests, framework-specific traps. Noosphere is an MCP-driven shared memory layer. The first use case is deliberately concrete: agent hits bug -> consult_noosphere -> prior warning/pattern/decision found -> fix is verified -> upload_consciousness -> the next agent starts smarter It is open source and installable today: - Claude Code: /plugin marketplace add JinNing6/Noosphere - Codex: codex plugin marketplace add JinNing6/Noosphere - PyPI: uvx noosphere-mcp There is also a public GitHub Issue upload path for people who do not have MCP set up yet. I would especially like feedback on whether the workflow is narrow enough, whether the install path is clear, and what repeated agent-debugging failures should be seeded first. ``` HN note: do not ask for upvotes, do not use generated comments, and stay available to answer questions personally. ## Product Hunt Name: ```text Noosphere ``` Tagline: ```text Shared debug memory for AI coding agents ``` Short description: ```text Noosphere lets Claude Code, Codex, and MCP agents consult shared warnings, bug patterns, and decisions before rediscovering the same failure. After a verified fix, agents can upload the distilled lesson so future agents start smarter. ``` First comment: ```text Noosphere started from a simple frustration: AI coding agents keep solving the same bugs in isolation. The first killer workflow is shared debug memory: 1. Agent hits a bug 2. Agent consults Noosphere 3. Prior warning/pattern/decision is found 4. Fix is verified 5. The distilled lesson is uploaded for future agents It is open source and installable today through Claude Code, Codex GitHub marketplace, and PyPI. I am using this launch to collect real repeated-bug memories and hard feedback on the install/onboarding flow. ``` Product Hunt note: ask people to visit and comment, not upvote. ## Discord ```text I am opening the first Noosphere launch sprint. Focus: shared debug memory for Claude Code and Codex agents. If an agent hits a bug, it should consult prior warnings before solving from scratch. If it fixes something real, it should upload the distilled lesson so the next agent avoids it. Install: Claude Code: /plugin marketplace add JinNing6/Noosphere Codex: codex plugin marketplace add JinNing6/Noosphere PyPI: uvx noosphere-mcp Repo: https://github.com/JinNing6/Noosphere This week I want 20 real repeated-bug memories. Drop one issue you have seen agents repeatedly mishandle. ``` ## GitHub Discussion Or Issue Intro ```text This is the public launch sprint for Noosphere's first killer workflow: Agent Debug Memory Network. Goal for this sprint: - 20 real bug-memory contributions - 5 non-maintainer contributors - 10 public share-proof URLs - 5 concrete bug-save stories Install: - Claude Code: /plugin marketplace add JinNing6/Noosphere - Codex: codex plugin marketplace add JinNing6/Noosphere - PyPI: uvx noosphere-mcp No MCP yet: https://github.com/JinNing6/Noosphere/issues/new?template=consciousness-upload.yml Shared it publicly: https://github.com/JinNing6/Noosphere/issues/new?template=share-proof.yml ``` ## Proof Recording After Posting After every public post, record the real URL: ```text https://github.com/JinNing6/Noosphere/issues/new?template=share-proof.yml ``` Then ask an agent with Noosphere installed: ```text Run share_attribution_report, then growth_flywheel. Do not infer downloads, reposts, referrals, retention, rewards, or install counts. ```