Weekly drop 2026-07-25: 28 added, 0 updated - #6
Merged
Conversation
Mode A (launched/updated 2026-07-10 → 2026-07-24): Screenpipe, Buzz, Palmier Pro, OneCLI, Orca, OpenViking, Rowboat, Natural, CartAI, Crunchbase MCP, Concho, DeepSQL, Nobie, Echo, CodeAlmanac, Lyzr Mode B (established gaps): AutoGPT, LocalAI, promptfoo, Cognee, Graphiti, DeepEval, Arize Phoenix, Genkit, Helicone, Postgres MCP Pro, Magic MCP, mcp-agent All star counts verified live via gh api; all URLs verified loading. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Window: last drop was
2026-07-17, so the lookback is Jul 17 → Jul 24 plus the mandatory 7-day gap sweep back to Jul 10.Catalog: 303 → 331 agents, 20 categories,
validate-content.tspasses with 0 violations.Verification: every star count below pulled live via
gh api repos/<owner>/<repo>today (never memory); every non-GitHub URL fetched with a browser UA and confirmed200. No repo in this drop is archived.✅ TO ADD — Mode A (launched / majorly updated in window + gap sweep)
Screenpipe —
screenpipe· repo · 20,523★ (verified), pushed 2026-07-25Signal: Launch HN: Screenpipe (YC S26) — 84 pts, 2026-07-23. YC S26. 24/7 local screen+audio capture served as context to Claude/Codex.
cli,api· featuredBuzz —
buzz· repo · 11,325★ (verified), pushed 2026-07-25Signal: HN front page 376 pts, 2026-07-21 + TechCrunch. Built by Block. Nostr-based workspace = team chat + Git forge + agents with cryptographic identities. Apache 2.0.
cli,api· featuredPalmier Pro —
palmier-pro· repo · 12,090★ (verified), pushed 2026-07-25Signal: Show HN — 187 pts, 2026-07-23. YC S24. Real MCP surface (HTTP MCP at
127.0.0.1:19789/mcp), so agents drive the video timeline — not GUI-only.mcpOneCLI —
onecli· repo · 2,813★ (verified), pushed 2026-07-25Signal: Show HN: OSS credential gateway that keeps secrets out of AI agents — 103 pts, 2026-07-23. Injects API keys at request time so agents never see plaintext.
cli,api,mcpOrca —
orca· repo · 28,839★ (verified), pushed 2026-07-25Signal: +466 stars the week of Jul 12; near-daily releases. MIT. Runs fleets of Claude Code/Codex/OpenCode in parallel git worktrees.
cli,api· featuredOpenViking —
openviking· repo · 27,209★ (verified), pushed 2026-07-25Signal: backed by Volcengine (ByteDance cloud); trending in window. Context DB exposing memories/skills as a
viking://virtual filesystem with tiered on-demand loading.cli,api,mcpRowboat —
rowboat· repo · 16,790★ (verified), pushed 2026-07-24Signal: YC S24, trending in window. Desktop coworker building a backlinked knowledge graph over email/meetings/Slack; MCP-extensible.
mcp,apiNatural —
natural· natural.com (200 ✓)Signal: $30M Series A led by Forerunner Ventures, announced 2026-07-20 (TechCrunch, PYMNTS, Finextra); ~$40M total. Agent payment rails — wallets, vaults, agent-to-agent transfers. Ships Python SDK (
naturalpay) + MCP server.api,mcpDistinct from existing
hilt-pay,sequence,bankr,brex-ai.CartAI —
cartai· cartai.ai (200 ✓)Signal: Product Hunt front-page launch 2026-07-21 (SaaS/Dev Tools/AI). Commercial-signal track. Agentic checkout on live merchant surfaces over Visa/Mastercard agent rails; dev API + open-source MCP server.
api,mcpCrunchbase MCP —
crunchbase-mcp· about.crunchbase.com (200 ✓)Signal: first-party MCP announced 2026-07-21 (Yahoo Finance, AccessNewswire); listed on Glama + LobeHub. 39B private-market signals, 790 fields/company.
mcp,apiDistinct from the older third-party
Cyreslab-AI/crunchbase-mcp-server.Concho —
concho· concho.ai (200 ✓)Signal: launched 2026-07-14, covered by SiliconANGLE + GlobeNewswire. "Cognitive precompiler" that turns an enterprise codebase into a queryable knowledge layer served over MCP.
mcpDeepSQL —
deepsql· deepsql.ai (200 ✓)Signal: Show HN: A self-hostable DBA agent for Postgres and MySQL — 52 pts, 2026-07-20. Query optimization, index advice, health monitoring; in-VPC self-host.
mcp,cliNobie —
⚠️ Listed
nobie· nobie.com (200 ✓)Signal: Show HN: an Excel-compatible runtime for agents and humans — 96 pts, 2026-07-13 (gap sweep). Identifiable company (SF/London/Toronto).
clicli-only — I could not confirm a public HTTP API from the landing page. Flagging rather than overstating.Echo —
echo· echo.tracerml.ai (200 ✓)Signal: 463 pts on HN, 2026-07-23 — "Fable-level results at 1/3 the cost using open-weight models." Tracer (YC-backed). Routes each request across specialist open-weight models behind one endpoint.
api, streamingComplements existing
workweave-router.CodeAlmanac —
⚠️ Weakest access story in the drop — CLI-only, explicitly no MCP/API yet. Included on YC + Show HN signal; happy to drop if you'd rather hold it.
codealmanac· repo · 636★ (verified), pushed 2026-07-25Signal: Show HN: Karpathy-style codebase wiki from your conversations — 60 pts, 2026-07-21. YC S26, 843 commits, active PRs.
cliLyzr —
⚠️ Date edge: the raise was announced 2026-07-09, one day before the gap-sweep window opens, though the news cycle ran Jul 9–13. Your call — this is the only entry that sits outside the window proper.
lyzr· lyzr.ai (200 ✓)Signal: $100M Series B at ~$500M valuation (TechCrunch, Bloomberg, TNW); deployed at Willis Towers Watson, VeriFone, US agencies. Sovereign/on-prem enterprise agent infra.
api✅ TO ADD — Mode B (established, missing from catalog)
autogpt· 185,679★ (verified), pushed 2026-07-25 · Significant Gravitas ·api,cli· featuredThe most glaring gap in the catalog:
babyagiandautogenare listed but the single most iconic autonomous-agent project was not.localai· 47,845★ (verified), pushed 2026-07-25 ·api,cli, streamingPeers
vllm/ollama/jan/lm-studioare all listed; LocalAI was the omission.promptfoo· 23,581★ (verified), pushed 2026-07-25 · VC-backed ·cli,apiLLM eval + AI red-teaming. Fills a thin spot:
semgrep/snykare code-focused, nothing covered prompt-injection/jailbreak scanning.cognee· 29,305★ (verified), pushed 2026-07-25 · Topoteretes ·api,mcp,cligraphiti· 29,191★ (verified), pushed 2026-07-25 · Zep AI ·api,mcpNote:
zepis already listed, but Graphiti is a separate, much-higher-star product line (temporally-aware graphs). Added per the "same company, new product" rule.deepeval· 17,112★ (verified), pushed 2026-07-24 · Confident AI ·cli,apiarize-phoenix· 10,728★ (verified), pushed 2026-07-25 · Arize AI ·api,clilangfuse/langsmith/datadog-llm-observabilitywere listed; Phoenix was the OSS gap.genkit· 6,283★ (verified), pushed 2026-07-25 · Google ·api,cli, streaminggoogle-adkwas listed; Genkit is a distinct first-party Google framework.helicone· 5,993★ (verified), pushed 2026-07-25 · YC W23 ·apipostgres-mcp-pro· 3,116★ (verified), pushed 2026-01-22 · Crystal DBA ·mcpListed across Glama/Smithery/PulseMCP/LobeHub. ClickHouse/Redis/Mongo/Neon/Supabase MCPs were listed but no Postgres equivalent.
magic-mcp· 5,517★ (verified), pushed 2026-07-24 · 21st.dev ·mcpmcp-agent· 8,478★ (verified), pushed 2026-01-25 · LastMile AI ·api,mcpCatalog had many MCP servers but few MCP-native agent frameworks.
🔄 UPDATED
None. I checked the live files for candidates that overlap existing entries (
zepvs Graphiti, the payments cluster vs Natural,crunchbase-mcp-servervs first-party) and found no field I could verify as factually wrong or outdated. I'd rather propose zero updates than invent diffs.⏭ REJECTED
💡 CATEGORY PROPOSALS
None — the 3+ candidate threshold was not met for a new category, so I mapped everything onto live slugs.
That said, one pattern is worth noting for a future run: LLM evaluation / testing now has 3 entries wedged into
code-devtools+data-analytics(promptfoo, DeepEval, Arize Phoenix), and the held-back Ragas would make 4. If you want anevaluation-testingcategory, next run is the moment — I did not create one here, per the rules.🤖 Generated with Claude Code