ChronoVec

Vector memory for data that changes

CI PyPI License

Most vector libraries optimize for a static corpus. Real systems do not stay still: agents write memories continuously, documents get corrected, users invoke the right to erasure, and teams need to reproduce what a retriever saw last Tuesday.

ChronoVec is an approximate-nearest-neighbour index built for that case. Every record carries a version interval, so queries can read the index as of any past moment, deletions physically reclaim space on a bounded budget, and continuous writes never force a rebuild — or block a reader.

pip install chronovec

The same native core is available through the other mainstream package managers once published:

cargo add chronovec
npm install @chronovec/native

Building from source requires a C++20 compiler and CMake — pip install . handles that automatically. See Getting started for full instructions.

Portable builds use baseline CPU instructions. If every deployment CPU supports AVX2, opt in with -DCHRONOVEC_ENABLE_AVX2=ON for faster x86 distance kernels.

Why ChronoVec exists

1. Query the past

t1 = index.insert(1, embedding)
index.delete(1)

index.search(query, k=10)               # now: id 1 is gone
index.search(query, k=10, snapshot=t1)  # as of t1: id 1 is there

Audit, reproducible evaluation, and retrieval regression debugging all need this. ChronoVec makes the snapshot part of the retrieval API.

2. Delete for real, on a budget

index.delete(user_vector_id)
index.vacuum(oldest_snapshot=index.clock + 1, budget_versions=64)

vacuum physically reclaims at most budget_versions expired records — no full rebuild. For compliance workflows, logical deletion is immediate and physical reclamation is explicit and bounded.

3. Write continuously without blocking readers

Writes never block reads. A reader snapshots an immutable view and queries it lock-free, even while a batch is in progress. Reads and writes compose: insert, delete, and query concurrently from multiple threads with no coordination at the reader.

4. Branch and speculate

from chronovec import AgentMemory

memory = AgentMemory(384)
memory.add(1, embedding, text="base knowledge")

plan = memory.branch("hypothesis")
plan.add(2, other, text="speculative — not yet committed")
plan.search(query)    # sees base + speculation
memory.search(query)  # main never saw the speculation
plan.discard()        # abandon speculation; purge later

Agent memory with snapshot isolation: spawn a branch, explore speculatively, discard or merge. Each branch gets its own private delta index by default, so there's no fixed branch-count ceiling.

5. Survive a crash

index = Index(768, wal_path="index.wal")
index.insert_many(ids, vectors)
# process dies here —
index = Index(768, wal_path="index.wal")  # replays the log; nothing lost

Writes are logged before they are applied. A torn tail from a crash mid-append is detected by CRC, dropped, and the file truncated cleanly.

Quick start

Quickstart

from chronovec import Collection

# In-memory collection, for easy prototyping. Add persistence easily below!
collection = Collection(dimensions=3)

# Add records in batches; update() and delete() use the same ids-oriented API.
collection.add(
    ids=["doc1", "doc2"],   # unique per record
    embeddings=[[1, 0, 0], [0, 1, 0]],  # or pass embedding_function= and add documents only
    documents=["This is document1", "This is document2"],
    metadatas=[{"source": "notion"}, {"source": "google-docs"}],  # filter on these!
)

# Query the 2 most similar results. get() fetches by id/filter without a search.
results = collection.query(
    [1, 0, 0],
    k=2,
    # where={"source": "notion"},   # optional metadata filter
    # snapshot=collection.snapshot(),  # optional: read as of an earlier point in time
)

For text embedding, pass either the original batch callable or ChronoVec's provider-neutral adapter. The adapter keeps document and query embedding separate when your model needs different instructions:

from chronovec import Collection, CustomEmbedding

embedder = CustomEmbedding(
    embed_documents=my_embed_documents,  # list[str] -> list[list[float]]
    embed_query=my_embed_query,           # str -> list[float]
)
collection = Collection(dimensions=384, embedding_function=embedder)
collection.add(ids=["doc1"], documents=["This is document1"])
results = collection.query(query_text="find document one", k=1)

Objects exposing embed_documents and embed_query are accepted directly, as are LlamaIndex-style get_text_embedding_batch and get_query_embedding objects. ChronoVec does not install or select a model provider; use the provider library that fits your application.

For local Sentence Transformers models, install the optional adapter:

pip install "chronovec[sentence-transformers]"

from chronovec import Collection, SentenceTransformerEmbedding

embedding = SentenceTransformerEmbedding("sentence-transformers/all-MiniLM-L6-v2")
collection = Collection(384, embedding_function=embedding)

For hosted providers, ChronoVec wraps LiteLLM so one adapter covers OpenAI, Cohere, Bedrock, Azure, and the rest of LiteLLM's provider list without ChronoVec depending on any of them directly:

pip install "chronovec[litellm]"

Pass the provider secret explicitly from your environment:

import os

from chronovec import Collection, ProviderEmbedding

embedding = ProviderEmbedding(
    provider="openai",
    model="text-embedding-3-small",
    api_key=os.environ["OPENAI_API_KEY"],
)
collection = Collection(1536, embedding_function=embedding)

Add persistence with a Client, which opens (or creates) named collections that checkpoint themselves to disk after every mutation — no separate save call:

from chronovec import Client

client = Client("./data")  # omit the path for "./.chronovec"

# get_collection, list_collections, delete_collection also available!
collection = client.get_or_create_collection("docs", dimensions=3)
collection.add(ids=["doc1"], embeddings=[[1, 0, 0]], documents=["This is document1"])

Inspect a persistent store with chronovec list ./data or chronovec inspect ./data docs.

# Install from source — cmake runs automatically
pip install ".[dev]"

from chronovec import Index, Collection

# Low-level API: int64 ids, NumPy vectors
index = Index(384, metric="cosine", page_capacity=256, nprobe=16)
t1 = index.insert(1, embedding)
index.insert(2, other_embedding)

for hit in index.search(query, k=10):
    print(hit.id, hit.distance)

index.delete(1)
index.vacuum(oldest_snapshot=index.clock + 1, budget_versions=64)
index.save("memories.cvec")
restored = Index.load("memories.cvec")

# High-level API: string ids, metadata, rich filtering
memory = Collection(dimensions=768, metric="cosine")
memory.add(ids=["doc-a"], embeddings=[v], metadatas=[{"lang": "en", "score": 0.9}])

before = memory.snapshot()
memory.add(ids=["doc-a"], embeddings=[corrected])

memory.query(q, k=5, where={"lang": "en"})                  # latest
memory.query(q, k=5, where={"lang": "en"}, snapshot=before) # as of `before`

where supports $eq $ne $gt $gte $lt $lte $in $nin $contains $regex $and $or.

# Async API for FastAPI / asyncio — same shape as Collection, awaited
from chronovec import AsyncCollection

memory = AsyncCollection(768, embedding_function=my_embed)
await memory.add(ids=["doc-a"], documents=["the user prefers dark mode"])
results = await memory.query(query_text="appearance settings", k=5)

Framework integrations

Integration Import Notes
LangChain chronovec.integrations.langchain.ChronoVecVectorStore Full VectorStore subclass with branching for LCEL chains
LlamaIndex chronovec.integrations.llamaindex.ChronoVecLlamaStore Node storage contract, metadata filtering
LangGraph chronovec.integrations.langgraph.LangGraphMemory Maps one graph thread_id to one isolated ChronoVec branch
DuckDB chronovec.duckdb_adapter.ChronoDuckDBAdapter SQL UDF interface (experimental)
SQLite native virtual table Embedded, durable via SQLite WAL
Rust bindings/rust/chronovec Raw Index tier plus an ergonomic collection::Collection (string ids, metadata, filters — same vocabulary as Python's Collection)
Go bindings/go/chronovec Raw Index tier via cgo, plus an ergonomic Collection (any-typed ids, map[string]any metadata, where=-style filter maps) built from scratch — no Rust crate for Go to reuse
Node.js bindings/node (@chronovec/native) Raw Index tier plus an ergonomic Collection (string ids, metadata, filter DSL as plain where= objects), via napi-rs wrapping the Rust crate. ids are bigint (JS number can't hold the full id range losslessly) 
# LangChain — branching LCEL chains
from chronovec.integrations.langchain import ChronoVecVectorStore

store = ChronoVecVectorStore(embedding=embeddings, dimensions=384)
store.add_texts(["the user prefers dark mode"])

with store.branch("hypothesis") as scratch:
    scratch.add_texts(["speculative memory"])
    scratch.similarity_search("theme")  # sees both
store.similarity_search("theme")         # speculation discarded

# LlamaIndex
from chronovec.integrations.llamaindex import ChronoVecLlamaStore

store = ChronoVecLlamaStore(dimensions=768)
index = VectorStoreIndex.from_vector_store(store)

Language support

Language Status
C / C++ Native library and stable C ABI (bindings/rust/chronovec-sys/native/include/chronovec.h)
Python Stable — Client/Collection for applications, Index for low-level control
Rust Supported alpha — source-available, not yet published to crates.io; the native core builds from the crate
Go Supported alpha — github.com/mchl-labs/chronovec/bindings/go/chronovec via cgo
Node.js Supported alpha — source-available, not yet published to npm; builds platform binaries from source
SQLite Integration — loadable virtual table and static-registration library

See the support and maturity matrix for the exact compatibility and guarantee boundary.

Performance

Measured fresh against chromadb, faiss-ivfflat, and hnswlib on SIFT-128, GloVe-25, and GIST-960, up to 500,000 live vectors — ChronoVec leads on the workload it's built for, and the margin grows with scale rather than shrinking:

  • Write-heavy workloads: chronovec holds ~140–150k replacement-pairs/s at high recall across every churn epoch tested at 200k live vectors — 16.5x faiss's steady-state throughput at that scale (vs. 2x at 10k; faiss's remove_ids cost compounds with corpus size). hnswlib stays roughly 70x below chronovec throughout, and chroma didn't complete this workload above 10,000 live vectors — three independent attempts at 50k/100k/200k all exceeded their 5–24 minute time budgets.
  • Concurrent reads during writes: chronovec is the only one of the four that's both provably safe (snapshot isolation) and fast — hnswlib and faiss don't document concurrent read+write as safe at all, and chroma is safe but ~980x slower at p99 at 200,000 live vectors.
  • Metadata-filtered search: chronovec's page-skipping gets faster as the filter tightens, an advantage that holds from 50k through 500k live vectors; hnswlib gets up to 100x slower at tight filters, and faiss's post-filter approach is fast but falls short of its recall target by up to 80 points and, at scale, stops returning k results for the majority of queries.

For a corpus that's built once and never mutated afterward, hnswlib and faiss both out-query chronovec by ~1.5–2x at matched recall on SIFT-128/GloVe-25 — that's the price of the versioning machinery above, and if your index never changes, either is a fine choice. ChronoVec overtakes hnswlib again at 960 dimensions. ChronoVec is for the other workload: agents writing memories, RAG corpora receiving corrections, CDC streams, and systems that need deletion, historical replay, or filtered queries against a corpus that keeps changing.

See docs/performance.md for full numbers, methodology, and the regression gate protocol.

The public comparison suite is reproducible from a clean checkout: install the benchmark extra, fetch the named ANN-Benchmarks files, then run benchmarks/run_all.sh. The command writes machine-readable JSON results and skips only engines that are not installed; see the benchmark instructions.

Examples

Example What it shows
examples/agent_memory.py Branching, speculation, time travel, vacuum
examples/langgraph_branching_memory.py LangGraph-style trajectory evaluation with isolated ChronoVec memory branches
examples/lats_chronovec.py Persistent-delta LATS tree search with 85 live isolated trajectories and atomic publish
examples/lats_benchmark.py LATS at scale: configurable depth/branching-factor, async concurrent rollouts, checkpoint/resume, and a benchmark against a copy-per-trajectory baseline
examples/rag_time_travel.py Query corpus as of t-1 after an update
examples/streaming_updates.py High-churn loop: insert/delete/vacuum, amplification stays flat
examples/compliance_deletion.py GDPR erasure: delete + vacuum + verify gone at any snapshot

Flagship demo: agent memory

Run the offline demo—no model key, database, or service required:

pip install chronovec
python examples/agent_memory.py

It captures an interaction, retries a mistaken memory update on an isolated branch, and merges the corrected result without contaminating the original timeline. For the broader mutable-RAG workflow, see examples/mutable_rag.py.

Documentation
Getting started Install, build, 3-minute tour
Migrating from Chroma Familiar API, snapshots, branching, and deletion semantics
Architecture MVCC, page structure, routing, reclamation
API reference All public classes and methods
Integrations LangChain, LlamaIndex, DuckDB, SQLite, Rust
Use cases Agent memory, RAG with history, compliance, streaming
Performance Benchmarks, methodology, regression gates
Contributing Build from source, conventions, how to add tests

Using ChronoVec from a coding agent

If you are building or modifying code that uses ChronoVec inside Claude Code, a skill file is included that teaches the agent the API idioms, snapshot timing, branching patterns, and common footguns. It activates automatically when relevant. The skill lives at .claude/skills/chronovec/chronovec/SKILL.md and can be invoked explicitly with /chronovec.

Citing ChronoVec

If ChronoVec contributes to your research or publication, please cite it:

@software{chronovec,
  title  = {ChronoVec: A Versioned Vector Index with Snapshot Isolation},
  author = {Rottoli, Michael},
  year   = {2026},
  url    = {https://github.com/mchl-labs/chronovec},
  license = {Apache-2.0}
}

The repository also includes a machine-readable CITATION.cff file.

Status

  • Solid: MVCC correctness, bounded reclamation, lock-free reads, checkpoint save/load, sanitizer targets (ASan/UBSan/TSan), C-ABI fuzzing, comprehensive test suite (3,700+ lines).
  • Stable: The Python API and C ABI are production-supported. Writers serialize on a single MVCC lock (lock-free reads, not multiwriter mutation).
  • Supported alpha: Rust, Go, and Node.js bindings have release CI but do not yet carry the Python API's compatibility promise.
  • Experimental: DuckDB adapter (not thread-safe, minimal test coverage), SQLite dqlite failover.
  • Async: AsyncCollection is available for asyncio/FastAPI applications; the low-level Index can be called through asyncio.to_thread() when needed.

Project links

Data verified by PyPI on Sep 7, 2026
Data provided by the project maintainers, verified at the time the release was uploaded to PyPI.

Key dates

PyPI data
Data sourced directly from PyPI's database.
  • Released:
Latest release

1 maintainer

PyPI data
Data sourced directly from PyPI's database.
Avatar for mchl-labs from gravatar.com mchl-labs

Credits

Author: mchl-labs

GitHub Statistics

Data verified by PyPI on Sep 7, 2026
The GitHub source repository was provided by the project maintainers and verified by PyPI at the time of upload. Stars, forks, and open issues/PRs are derived from that repository and have not been independently verified.

License expression

Apache-2.0
View SPDX License List

Requires

Python >=3.11

Provides Extra

sentence-transformers litellm embeddings duckdb integrations langgraph test test-extra benchmarks dev

Tags

vector-search ann mvcc agent-memory rag vector-database

Classifiers

Report project as malware

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

chronovec-1.0.0.tar.gz (4.4 MB view details)

Uploaded Source

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

chronovec-1.0.0-cp313-cp313-win_amd64.whl (188.3 kB view details)

Uploaded CPython 3.13Windows x86-64

chronovec-1.0.0-cp313-cp313-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl (290.3 kB view details)

Uploaded CPython 3.13manylinux: glibc 2.24+ x86-64manylinux: glibc 2.28+ x86-64

chronovec-1.0.0-cp313-cp313-macosx_11_0_arm64.whl (183.4 kB view details)

Uploaded CPython 3.13macOS 11.0+ ARM64

chronovec-1.0.0-cp312-cp312-win_amd64.whl (188.3 kB view details)

Uploaded CPython 3.12Windows x86-64

chronovec-1.0.0-cp312-cp312-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl (290.3 kB view details)

Uploaded CPython 3.12manylinux: glibc 2.24+ x86-64manylinux: glibc 2.28+ x86-64

chronovec-1.0.0-cp312-cp312-macosx_11_0_arm64.whl (183.4 kB view details)

Uploaded CPython 3.12macOS 11.0+ ARM64

chronovec-1.0.0-cp311-cp311-win_amd64.whl (188.3 kB view details)

Uploaded CPython 3.11Windows x86-64

chronovec-1.0.0-cp311-cp311-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl (290.3 kB view details)

Uploaded CPython 3.11manylinux: glibc 2.24+ x86-64manylinux: glibc 2.28+ x86-64

chronovec-1.0.0-cp311-cp311-macosx_11_0_arm64.whl (183.4 kB view details)

Uploaded CPython 3.11macOS 11.0+ ARM64

File details

Details for the file chronovec-1.0.0.tar.gz.

File metadata

  • Download URL: chronovec-1.0.0.tar.gz
  • Upload date:
  • Size: 4.4 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for chronovec-1.0.0.tar.gz
Algorithm Hash digest
SHA256 d2c22d3887393251cd7c981dc51470dee2a02caeda53dd7e69066c57d3c9be5a
MD5 a5278df08f31d3b773436919679db015
BLAKE2b-256 8559f603b320b90232573d5558e8de18994c30228e6f139db6dda6e66ec6622a

See more details on using hashes here.

Provenance

The following attestation bundles were made for chronovec-1.0.0.tar.gz:

Publisher: publish.yml on mchl-labs/chronovec

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file chronovec-1.0.0-cp313-cp313-win_amd64.whl.

File metadata

  • Download URL: chronovec-1.0.0-cp313-cp313-win_amd64.whl
  • Upload date:
  • Size: 188.3 kB
  • Tags: CPython 3.13, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for chronovec-1.0.0-cp313-cp313-win_amd64.whl
Algorithm Hash digest
SHA256 16976aecd573b46b9224a09b8622bab8d14e5b094f444fe5b57c785ad447037f
MD5 5213470b0e10788600b2510133a28cc7
BLAKE2b-256 6714e903b87a59f813366282afebaaa784d516c9134818df96d023abef037e5e

See more details on using hashes here.

Provenance

The following attestation bundles were made for chronovec-1.0.0-cp313-cp313-win_amd64.whl:

Publisher: publish.yml on mchl-labs/chronovec

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file chronovec-1.0.0-cp313-cp313-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl.

File metadata

File hashes

Hashes for chronovec-1.0.0-cp313-cp313-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl
Algorithm Hash digest
SHA256 1c1f86c243da3361bd3a12fc65c26e7a756f96a31822da69dba0baff9f0d90c3
MD5 9e3898f5ab5ff34e7008865190f79436
BLAKE2b-256 633ece1063899bd5e847319fbe8641d97fe86706e055d9ff9bbe9feb489e55b0

See more details on using hashes here.

Provenance

The following attestation bundles were made for chronovec-1.0.0-cp313-cp313-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl:

Publisher: publish.yml on mchl-labs/chronovec

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file chronovec-1.0.0-cp313-cp313-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for chronovec-1.0.0-cp313-cp313-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 fcce19cdb834bbce7ca83223166e28e598c71fa8a1a66f67f64447add96c4546
MD5 2455e2b55af02b3c1e30a3232d5b2316
BLAKE2b-256 2a18d0eecdb4a49f6b40d01857a9bd3227b5be0a95a6b9fab690386f00d2f912

See more details on using hashes here.

Provenance

The following attestation bundles were made for chronovec-1.0.0-cp313-cp313-macosx_11_0_arm64.whl:

Publisher: publish.yml on mchl-labs/chronovec

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file chronovec-1.0.0-cp312-cp312-win_amd64.whl.

File metadata

  • Download URL: chronovec-1.0.0-cp312-cp312-win_amd64.whl
  • Upload date:
  • Size: 188.3 kB
  • Tags: CPython 3.12, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for chronovec-1.0.0-cp312-cp312-win_amd64.whl
Algorithm Hash digest
SHA256 a54c5dd69f91b0ec3fdca2102b43f1e6741305e003b3bc3bf39ae0827ccbd77a
MD5 b48c4bfe1fb1aee66d68a82a9629abbd
BLAKE2b-256 8ad01414f762450065b0d6478139930a07959736a5d32cae230ccde9ff655c51

See more details on using hashes here.

Provenance

The following attestation bundles were made for chronovec-1.0.0-cp312-cp312-win_amd64.whl:

Publisher: publish.yml on mchl-labs/chronovec

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file chronovec-1.0.0-cp312-cp312-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl.

File metadata

File hashes

Hashes for chronovec-1.0.0-cp312-cp312-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl
Algorithm Hash digest
SHA256 292b5f383329a2b3a2d9908232973f5b3c52e951eff9b3eb0f9f3aec6f087b0a
MD5 6385102ea2ea1f9607fd96a46fe91ec0
BLAKE2b-256 4a0289235b3282b6d868f4e0aecef75e6b51842c6db53ce988e84483021a7a21

See more details on using hashes here.

Provenance

The following attestation bundles were made for chronovec-1.0.0-cp312-cp312-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl:

Publisher: publish.yml on mchl-labs/chronovec

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file chronovec-1.0.0-cp312-cp312-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for chronovec-1.0.0-cp312-cp312-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 460578cc450385e14e93fee5952bbb042689db7a353102a166cf7112e83f846b
MD5 cd480e502861cd868fc57434e0c9040c
BLAKE2b-256 81e611226bff0c5ffecbe7e502f614183fb40cc7e1db00949e0025dd32062094

See more details on using hashes here.

Provenance

The following attestation bundles were made for chronovec-1.0.0-cp312-cp312-macosx_11_0_arm64.whl:

Publisher: publish.yml on mchl-labs/chronovec

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file chronovec-1.0.0-cp311-cp311-win_amd64.whl.

File metadata

  • Download URL: chronovec-1.0.0-cp311-cp311-win_amd64.whl
  • Upload date:
  • Size: 188.3 kB
  • Tags: CPython 3.11, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for chronovec-1.0.0-cp311-cp311-win_amd64.whl
Algorithm Hash digest
SHA256 220dc54b183391ba23ee88f0b9c1843e31b909aa5926f24126b3d5ea21d2bd91
MD5 6c74fb1e007924296378f8c779ddb0e6
BLAKE2b-256 e29c0f9b543e84a53756449c6694bd1dad079381c436690f83d8115c0b38d18f

See more details on using hashes here.

Provenance

The following attestation bundles were made for chronovec-1.0.0-cp311-cp311-win_amd64.whl:

Publisher: publish.yml on mchl-labs/chronovec

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file chronovec-1.0.0-cp311-cp311-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl.

File metadata

File hashes

Hashes for chronovec-1.0.0-cp311-cp311-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl
Algorithm Hash digest
SHA256 b35d6c19107b160712a2ab7d71d105977c6cae02fbafc4c2faef7cfe21ff7eb4
MD5 713ac9b4379251f357a0b2b9c8d9a42b
BLAKE2b-256 747c89e560902770d94edce9f9562e3cb48d70ab70e32b02c0f366774b5fdaf6

See more details on using hashes here.

Provenance

The following attestation bundles were made for chronovec-1.0.0-cp311-cp311-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl:

Publisher: publish.yml on mchl-labs/chronovec

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file chronovec-1.0.0-cp311-cp311-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for chronovec-1.0.0-cp311-cp311-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 82ef6b743c19d0d3f0ba0ac1732daa807d9343498996d7793e117978d7d6e0e1
MD5 928b90d79dc33b1fb5b8e6819433c277
BLAKE2b-256 8ccc970a3b4c96340b9d54636c52687e4ddfe4876931c556d667ebf834c4b625

See more details on using hashes here.

Provenance

The following attestation bundles were made for chronovec-1.0.0-cp311-cp311-macosx_11_0_arm64.whl:

Publisher: publish.yml on mchl-labs/chronovec

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Release history Release notifications | RSS feed

This release

1.0.0 This release

10 files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page