GCF Structured Data Format

A compact, schema-free, lossless wire format for structured model and agent communication.

Stars
★ 63
Last updated
1d ago
License
Apache-2.0

At a glance

How it runs
Library / SDKMCP serverAgent plugin / skill
Works with
Universal · cross-platformCodex · Claude Code
Cost
Free, no paid service needed
Setup effort
Low · running in minutes
You'll need
Shell / CLINetwork accessMCP Server
Typical use
MCP server maintainers who want smaller structured tool responses while retaining decodable types and structure.
Not a fit if
  • Teams required to keep JSON on the wire without an encode/decode step
  • Organizations requiring unambiguous license metadata before adoption
  • Teams needing a mature, stable .NET SDK
Website
gcformat.com ↗ set on GitHub
Maintenance
Commits in the last 30 dayslast commit 2026-11-20latest release v3.6.0 (2026-10-03)
Source review
65/100 · Some gaps

What does this agent do, and when should you use it?

GCF is a structured-data wire format for LLM and multi-turn agent interactions, not an autonomous task-running agent. Its generic profile encodes ordinary objects, arrays, and scalars, while its graph profile represents symbols, edges, and distance groups; data can be converted from and back to JSON, YAML, TOML, CSV, or MessagePack. The project supplies Go, TypeScript, Python, Rust, Swift, Kotlin, and .NET SDKs alongside CLIs, streaming encoders, an MCP proxy, Claude Code and Codex plugins, and editor integrations. For repeated interactions, local IDs, session deduplication, and delta encoding reduce retransmitted context, while streaming uses deferred counts and a summary trailer to emit rows incrementally. The repository reports more than 2,500 model evaluations, interoperability validation across 17 serialization formats, and over 43 billion lossless round-trips; these figures are project-reported.

An application passes deserialized objects, arrays, or scalars to the generic profile's encodeGeneric API—called encode_generic in the Python example. The encoder declares object-array fields once, emits pipe-separated positional rows, and can flatten fixed-shape nested objects into > path columns. For code intelligence and knowledge graphs, the graph profile's encode function accepts Payload, Symbol, and Edge values, replaces repeated identifiers in edges with local IDs such as @0, and produces sections such as targets, related, and edges. decode() restores the structured value and can convert it to JSON, YAML, TOML, CSV, or MessagePack. A streaming encoder can emit [?], process cursor or traversal rows incrementally, and finish with a ##! summary containing the final counts. Across turns, previously declared symbols may be sent as references and changed context as deltas; gcf-proxy adds bidirectional conversion, session deduplication, and an HTTP frontend around an existing MCP server without source changes.

  1. MCP server maintainers who want smaller structured tool responses while retaining decodable types and structure.
  2. Multi-agent workflow developers repeatedly handing the same structured context between models and seeking lower token use.
  3. Code-intelligence or knowledge-graph systems that need compact symbols, relationship edges, and distance groups.
  4. Services processing large database cursors or graph traversals that must emit rows without buffering the full result.
  5. Applications ingesting JSON, YAML, TOML, CSV, or MessagePack that want one model-facing representation.
  6. Claude Code or Codex teams wanting the documented plugins and session-statistics hooks.

Key setup facts

Pulled from the install and usage notes below. The project README has the current values.

Install / run
  • pip install gcf-python
  • npm install @blackwell-systems/gcf
  • pip install gcf-proxy

How do you install or deploy this agent?

Install the SDK for the implementation language. The supplied material does not state minimum language versions or credential requirements.

pip install gcf-python
npm install @blackwell-systems/gcf
go get github.com/blackwell-systems/gcf-go
cargo add gcf
dotnet add package BlackwellSystems.Gcf

To wrap an existing MCP server, install the proxy:

pip install gcf-proxy

The project also lists Swift Package Manager, Kotlin/JitPack, Claude Code and Codex plugins, and integrations for VS Code, JetBrains, Zed, n8n, and tree-sitter. The supplied material does not include complete copyable installation arguments for the plugins and several of those integrations.

How do you use this agent?

This first Python example encodes ordinary structured data with encode_generic and requires no API key:

from gcf import encode_generic

output = encode_generic({
    "employees": [
        {"id": 1, "name": "Alice", "department": "Engineering", "salary": 95000},
        {"id": 2, "name": "Bob", "department": "Sales", "salary": 72000},
        {"id": 3, "name": "Carol", "department": "Marketing", "salary": 85000},
    ],
})

Use the graph profile for code relationships:

from gcf import encode, Payload, Symbol, Edge

output = encode(Payload(
    tool="context_for_task", token_budget=5000, tokens_used=1847,
    symbols=[
        Symbol(qualified_name="github.com/org/repo/pkg.AuthMiddleware", kind="function", score=0.78, provenance="lsp_resolved", distance=0),
        Symbol(qualified_name="github.com/org/repo/pkg.NewServer", kind="function", score=0.54, provenance="lsp_resolved", distance=1),
    ],
    edges=[Edge(source="github.com/org/repo/pkg.NewServer", target="github.com/org/repo/pkg.AuthMiddleware", edge_type="calls")],
))

Use the SDK's decode() operation to restore data or output JSON, YAML, TOML, CSV, or MessagePack. The source states that a CLI is included with the six original language SDKs, but it does not provide the precise CLI invocation syntax here.

What are this agent's strengths and limitations?

Pros
  • One grammar covers general structured values and graph data with nodes, edges, and distance groups, with conversion paths for JSON, YAML, TOML, CSV, and MessagePack.
  • The project reports 50–92% fewer tokens than JSON, plus further reductions from local IDs, session deduplication, and delta encoding.
  • All seven language implementations commit to zero runtime dependencies beyond their standard libraries.
  • Deferred counts and a summary trailer support constant-memory row streaming for large cursors and graph traversals.
  • Documented integration paths include an MCP proxy, Claude Code and Codex plugins, and tooling for several editors and workflow environments.
Limitations
  • Adopters must add GCF encoding and decoding to their data path; systems that only understand established formats still require boundary conversion.
  • The supplied license information conflicts: repository metadata and the badge say Apache-2.0, while the implementation section and license heading say MIT.
  • The comprehension, token-efficiency, and TOON corruption figures are reported by the project; no independent replication is included in the supplied evidence.
  • Component maturity varies: the listed .NET release is 0.1.0, while most other SDKs are 2.5.2 or later.
  • Session deduplication and delta encoding require local IDs and persistent multi-turn state, adding lifecycle management beyond independent payload encoding.

How does this agent compare with similar options?

Compared with JSON, GCF removes repeated field names through shared headers, positional rows, and local IDs; the project reports 50–92% token savings, at the cost of an additional codec boundary. Unlike Protobuf, it requires no .proto schema and remains directly model-readable; unlike MessagePack, it is not binary. The project positions TOON as its main compact-text alternative and reports that GCF wins token efficiency on 15 of 16 datasets and reaches 91.2% comprehension on complex graphs versus 68.8% for TOON. These comparative figures come from the project's own evaluations.

Key facts side by side with the most closely related agents.

Agent Source review Form / cost Stars Updated Language Full support on
GCF Structured Data Format This agent 65 · Some gaps Library / SDKFree ★ 63 1d ago — Codex · Claude Code
LeanCTX Context Intelligence 69 · Some gaps CLIFree + model costs ★ 4k 4d ago Rust Codex · Claude Code · OpenAI API · Claude API
MCP Memory Service 68 · Some gaps MCP serverFree ★ 2.1k 1d ago Python ChatGPT · Codex · Claude Code · Claude.ai
Crush Terminal Coding Assistant 65 · Some gaps CLIFree + model costs ★ 29k today Go OpenAI API · Claude API

How does FollowAgents rate this agent?

FollowAgents source review · FARS-2.1
Some gaps
65/ 100 5-point scale 3.3 / 5
Trust 16/29
Reliability 8/14
Adaptability 14/18
Convention 13/18
Effectiveness 10/13
Verifiability 4/8
Why each dimension lost points
Trust16 / 29 · 2.8/5

The repository describes encode/decode boundaries and MCP proxy use, claims no runtime SDK dependencies, and the sampled CI workflow limits permissions to contents: read; Apache-2.0 licensing and project attribution are clear. Deductions: the sample does not explain confirmation, sensitive-data handling, or rollback, and the zero-dependency claim coexists with development dependencies in the root package.json, so full marks are not supported.

Reliability8 / 14 · 2.9/5

The repository provides a specification, conformance fixtures, a coverage-matrix gate, and a cross-language differential-fuzz workflow; some failures are reported through nonzero exits. Deduction: README, FAQ, and Getting Started disagree on spec versions and SDK versions/round-trip counts. CI configuration shows test mechanisms, but this static review did not execute them or confirm results.

Adaptability14 / 18 · 3.9/5

The docs name concrete scenarios including MCP responses, APIs/databases, code graphs, RAG, and multi-turn sessions, and identify cases where a scalar or JSON consumer should use another path. Deductions: environment and trigger guidance relies on broad fit claims and promotional absolutes; the sample does not adequately define model or runtime selection boundaries.

Convention13 / 18 · 3.6/5

The docs are well organized and include multi-language installation, code examples, an FAQ, and limitations; the license text is complete. Deductions: spec and implementation versions conflict across README, FAQ, and installation instructions; the sample does not show a clear changelog or detailed maintainer/update commitments.

Effectiveness10 / 13 · 3.8/5

Examples show compact output, decode interoperability, streaming, and multi-turn deduplication/deltas, which could reduce context tokens and support downstream consumers if the claims hold. Deductions: benefit and performance figures are primarily repository-reported summaries, and practical tradeoffs and costs are only lightly covered.

Verifiability4 / 8 · 2.5/5

The README points to benchmarks, papers, the specification, and evaluation results; CI files describe conformance coverage and cross-SDK differential fuzzing. Deductions: the sample omits raw evaluation logs and full benchmark data needed to verify individual claims; version and count inconsistencies appear across docs, and factual statements, inference, and promotional conclusions are not consistently separated.

Risks and how to mitigate them
  • The docs disagree on specification versions, SDK versions, and verification counts; check the target revision's spec and the exact SDK release before adopting.
  • Benchmark, round-trip, and model-performance figures are primarily repository claims; this review did not run tests or independently verify external papers or data.
Evidence confidence: Low Reviewed Nov 08, 2026 Reviewed revision 616b25f79184 New commits since this review; the score may not cover them
See the full review method →
Maintain this project? Add the FARS badge to your README

The badge shows the current review score and verdict and links back to this page. It refreshes about once an hour after a re-review.

FARS

[![FARS review](https://followagents.com/badge/gcf.svg)](https://followagents.com/en/agents/gcf)
<a href="https://followagents.com/en/agents/gcf"><img src="https://followagents.com/badge/gcf.svg" alt="FARS review"></a>

Recent releases

Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.

  1. v3.6.0 gcf v3.6.0 — constant-column factoring and value-grouping
    Release notes excerpt
    Tabular data repeats the same value down a column all the time: a region, a tenant, a status, a currency. v3.6.0 stops paying for that repetition on the wire, and banks the savings only where it was measured not to hurt how accurately a model reads the data back. Constant-column factoring (SPEC 7.4.7). A column holding one value across every record is written once in the header (region=us-east) instead of once per row. Automatic, no configuration, lossless. The saving grows with how wide the repeated value is and how many records carry it. …
    View on GitHub ↗
  2. v3.5.3 GCF Spec v3.5.3 — canonical int64 numeric domain
    Release notes excerpt
    v3.5.3 specifies the canonical numeric domain (Section 2.3.2): signed 64-bit integers (int64, [-2^63, 2^63-1]) for integer values and IEEE-754 double for non-integer values — the largest integer domain representable losslessly in every reference language. Earlier versions left the numeric domain implementation-defined. v3.5.3 makes it a specified, symmetric conformance property: - A decoder parses an integer literal into an exact int64-capable type and returns an out-of-range error outside int64. - An encoder returns an out-of-range error for a host integer outside int64. …
    View on GitHub ↗
  3. v3.5.2 v3.5.2: number-grammar digits are ASCII only (errata)
    Release notes excerpt
    Spec errata: number-grammar digits are ASCII only (Sections 2.3, 2.4) The number grammar's DIGIT rules are ASCII 0-9 per RFC 5234; non-ASCII Unicode decimal digits (Arabic-Indic U+0660-U+0669, fullwidth U+FF10-U+FF19, etc.) are not digits. A token such as 1.٥ (ASCII 1, ., then U+0665) does not match the number grammar and decodes as the string "1.٥", never the number 1.5. …
    View on GitHub ↗

FAQ

Does GCF require a model API or paid service?
No. The codecs run as local language libraries, and all seven SDKs claim standard-library-only runtime dependencies. Any model charges depend on the external service an adopter chooses to pair with the format.
Can it replace JSON everywhere?
It can serve as the wire format at the model boundary and decode back to JSON or other documented formats. Interfaces that only accept JSON still need conversion; existing schemas and validators can operate on decoded output.
Is MCP mandatory?
No. The language SDKs and included CLIs can be used directly. gcf-proxy is an optional integration for wrapping existing MCP servers.
How does streaming work when the row count is unknown?
The encoder emits [?] in the header, streams rows as they arrive, and closes with a ##! summary containing the actual counts.
Is the license unambiguous?
No. The supplied repository metadata and badge identify Apache-2.0, but two README sections state MIT. Prospective adopters should verify the repository's actual LICENSE file and obtain maintainer clarification.
Is GCF Structured Data Format safe to use?
FollowAgents' FARS review (Nov 08, 2026) scored GCF Structured Data Format 16/29 on trust and safety. None of the eight safety controls the review checks for was missing outright. The repository has changed since that review.
Is GCF Structured Data Format still maintained?
Yes. The latest commit on GitHub is from 2026-11-20, within the last 30 days. The latest release is v3.6.0 (2026-10-03).
Who makes GCF Structured Data Format?
GCF Structured Data Format is published on GitHub by the account blackwell-systems. FollowAgents has not verified which person or company is behind that account.
What is GCF Structured Data Format's official website?
The project's GitHub repository lists http://www.gcformat.com/ as its website. That link is set by the repository owner; the source code itself is on GitHub (blackwell-systems/gcf).
Was this page helpful?
View on GitHub ↗ Install ↓

Compare agents like this one

The same FARS review applied across the shortlist this agent qualifies for.

Related concepts