Repository navigation
Releases: agntn/archives
Releases · agntn/archives
Release list
v0.7.4
Immutable
release. Only release title and notes can be modified.
v0.7.4 is the next minor release.
👉 Changelog
🚀 Enhancements
🔥 Performance
- tools: list failures once (#184)
🩹 Fixes
- providers: quiet retries (#188)
- providers: keep capture URLs as recorded (#187)
- tools: type input errors (#186)
- providers: quiet expected 404s (#177)
💅 Refactors
- lift extensions out (#192)
📦 Build
🏡 Chore
🤖 CI
- release through uppt (#182)
❤️ Contributors
v0.7.3
v0.7.2
🚀 Enhancements
- Dig the archives over MCP without Node (#151)
- Decode Common Crawl in plain JavaScript (#155)
- Read Hungary's web archive too (#159)
- List New Zealand's web archive (#161)
🩹 Fixes
- docs: Plug the IPv6 hole in the rate limit (#157)
- Export the setConfigCwd the docs promise (#158)
- providers: Ask Common Crawl the right crawl (#162)
- docs: Keep previews off production data (#163)
💅 Refactors
- Hash cache keys without ohash (#160)
❤️ Contributors
v0.7.1
🚀 Enhancements
- content: Save a capture's bytes to a file (#138)
- Add Vefsafn provider (#139)
- Print CDX digest, status and size (#147)
- Read archives without a filesystem (#148)
🔥 Performance
- tools: Print each original URL once (#142)
🩹 Fixes
- docs: Skip Archive.today on the worker (#136)
- providers: Drop only the default port (#141)
- docs: Build on Node.js 26 (#145)
💅 Refactors
- Fetch archives without ofetch (#137)
❤️ Contributors
v0.7.0
🚀 Enhancements
- content: Report the SHA-256 of each body (#127)
🩹 Fixes
- content: Read the exact URL, not its alias (#126)
- release: Keep oxfmt off the changelog (#128)
- docs: Drop menu links to a missing /mcp (#129)
- docs: Stop asking to build extensions (#133)
💅 Refactors
📦 Build
🏡 Chore
- Apply automated updates (f35e8e5)
⚠️ Breaking Changes
- tools:
⚠️ Share one definition per tool (#130)
❤️ Contributors
v0.6.0
🚀 Enhancements
- snapshots: Look up several targets at once (#99)
- mcp: A restart is enough after src edits (#102)
- docs: Every page on the instrument shell (#106)
🩹 Fixes
- Fail fast on invalid processing limits (#72)
- Keep archived redirects from going live (#73)
- Compose request deadlines with cancellation (#82)
- diff: Skip the digest check on a first slice (#91)
- pi: Mark failed archive reads as errors (#93)
- docs: Build without the parent dist (#97)
- archive-today: Read a 404 timemap as empty (#96)
- content: Keep archived redirect targets (#100)
- wayback: Read pinned captures without CDX (#101)
- docs: Make the query limit hold (#114)
- diff: Show both sides of a refused pair (#115)
- pi: Catch up with Pi 0.99 (#116)
- wayback: Stop timing out on a slow index (#117)
- providers: Give slow archives their minute (#119)
- errors: Say why a fetch failed (#122)
📖 Documentation
- readme: Open with example.com in 2002 (#107)
📦 Build
- Fold four tool configs into Vite+ (#105)
🏡 Chore
⚠️ Require Node.js 26 (#104)
⚠️ Breaking Changes
⚠️ Require Node.js 26 (#104)
❤️ Contributors
v0.5.4
v0.5.3
v0.5.2
🩹 Fixes
- Deduplicate concurrent provider loads (#51)
- Avoid wildcarding exact archive URLs (#53)
- Remove the fake Common Crawl latest index (#56)
- Preserve Archive.today capture URLs (#57)
- Don't drop providers added during resolution (#58)
- Reject enum values outside tool schemas (#59)
- Stop promising raw archive bytes (#60)
- Restore provider labels on archive results (#61)
💅 Refactors
- Make shared Ox checks real (#55)
❤️ Contributors
v0.5.1
🚀 Enhancements
- Add native OMP extension (#25)
- Add Archive-It provider (#26)
- Serve archives over MCP (#28)
- Read archived page content (#29)
- Add Conifer provider (#27)
- Filter snapshot listings by time window (#33)
- Fetch Archive.today capture bodies (#42)
- Search Memento archives through MemGator (#46)
- Keep numeric limits visible to agents (#49)
- Clarify snapshot discovery routing (#50)
🔥 Performance
🩹 Fixes
- Cancel archive requests with tool calls (#30)
- Fall back to snapshot stamp for unreadable memento dates (#31)
- Forward the cancellation signal to Conifer searches (#32)
- Keep the OMP loader imports literal (#34)
- Build before type-checking the extensions (#36)
- List the newest Archive.today capture (#39)
- Treat the Common Crawl no-captures 404 as empty (#41)
🤖 CI
- Publish tagged releases through OIDC (#48)