We first define the nerve and then characterize which simplicial animae arise in this way.

Definition 25.1.1 (Nerve). The nerve of an \(\infty \)-category \(C\) is the simplicial anima \(N(C) \colon \simp \catop \to \An \) defined by \(N(C)_n := \Map ([n],C)\), i.e. as the composite \[ \simp \catop \hookrightarrow \Cat \catop \hookrightarrow \Cat _{\infty }\catop \xrightarrow {\Hom _{\Cat _{\infty }}(-,C)} \An . \] By Remark 1.8.4, this construction is functorial in \(C\), giving a functor \(N\colon \Cat _{\infty } \to \sAn \).

We may think of points of \(\Map ([n],C)\) as ‘strings of \(n\) composable morphisms in \(C\)’:

  • For \(n = 0\), \(\Map ([0],C) = \Fun (*,C)^{\simeq } \simeq C^{\simeq }\) is just the groupoid core of \(C\);
  • For \(n = 1\), \(\Map ([1],C)\) is the anima of morphisms in \(C\);
  • For \(n = 2\), \(\Map ([2],C)\) is the anima of commutative triangles in \(C\), which by the Segal axiom, Proposition 1.4.5, is equivalent to \(\Map ([1],C) \times _{C^{\simeq }} \Map ([1],C)\);
  • More generally, the ‘higher version’ of the Segal condition inductively implies that for every \(n \geq 0\) we have \[ \Map ([n],C) \iso \Map ([1],C) \times _{C^{\simeq }} \Map ([1],C) \times _{C^{\simeq }} \dots \times _{C^{\simeq }} \Map ([1],C) \] for all \(n\).

The nerve records the objects of \(C\) in \(N(C)_0=C^{\simeq }\) and its hom animae as the fibers of \((d_1,d_0)\colon N(C)_1\to N(C)_0\times N(C)_0\). Its \(2\)-simplices record composition, and the higher simplices record the coherences. Theorem 25.1.10 will show that the nerve determines \(C\).

The Segal condition allows us to interpret a simplicial anima in the same way, with objects, morphisms, and composition:

Definition 25.1.2 (Segal anima). A simplicial anima \(X\colon \simp \catop \to \An \) is called a Segal anima if the Segal condition is satisfied: For every \(n \geq 1\), the inclusion maps \(e_i\colon [1] \cong \{i- 1\leq i \} \hookrightarrow [n]\) induce an isomorphism of animae \[ (e_1^*, \dots , e_n^*)\colon X_n \iso X_1 \times _{X_0} X_1 \times _{X_0} \dots \times _{X_0} X_1. \] We write \(\Seg (\An ) \subseteq \sAn = \Fun (\simp \catop ,\An )\) for the full subcategory of Segal animae.

The nerve \(N(C)\) of an \(\infty \)-category \(C\) is always a Segal anima, see Chapterexercise 25.2.

One may show, just like we did for \(\infty \)-categories in Section 1.4.2, that a Segal anima \(X\) comes equipped with a composition map \[ - \circ -\colon X_1 \times _{X_0} X_1 \simeq X_2 \xrightarrow {d_1} X_1 \] and that composition is unital and associative.

Convention 25.1.3. Given a Segal anima \(X\), we refer to \(0\)-simplices \(x \in X_0\) as objects of \(X\), and to 1-simplices \(f\in X_1\) as morphisms of \(X\). The notation \(f\colon x \to y\) includes specified identifications \(d_1(f)=x\) and \(d_0(f)=y\) in \(X_0\). Given objects \(x,y \in X_0\), we define the hom anima in \(X\) as the following pullback:

Commutative square with maps \Hom X(x,y) to X 1, * to X 0 times X 0 labeled (x,y), \Hom X(x,y) to *, X 1 to X 0 times X 0 labeled (d 1,d 0).

A \(2\)-simplex with boundary \(f\colon x\to y\), \(g\colon y\to z\), and \(h\colon x\to z\) supplies an identification \(g\circ f=h\) in \(\Hom _X(x,z)\).

Definition 25.1.4. Let \(X\) be a Segal anima. A morphism \(f\colon x\to y\) in \(X\) is called an isomorphism if there exist morphisms \(g,h\colon y\to x\) together with identifications \[ g\circ f=\id _x \quad \text {in }\Hom _X(x,x), \qquad f\circ h=\id _y \quad \text {in }\Hom _X(y,y). \] We write \(X_1^{\sim }\subseteq X_1\) for the full subanima spanned by the isomorphisms.

An identification between points \(x,y\in X_0\) induces an isomorphism from \(x\) to \(y\) in \(X\), by transporting the identity morphism. An isomorphism in \(X\) need not arise this way: for example, the ordinary simplicial nerve of a \(1\)-category, viewed as a simplicial anima with discrete terms, may have distinct isomorphic objects. The completeness condition below will ensure that identifications in \(X_0\) account for all isomorphisms in \(X\).

Definition 25.1.5 (Dwyer-Kan equivalences). A morphism \(f\colon X \to Y\) of Segal animae is called a Dwyer-Kan equivalence if the following two conditions are satisfied:

(1)

(Essential surjectivity): For every object \(y\in Y_0\) there exists an object \(x\in X_0\) and an isomorphism \(f(x)\cong y\) in the Segal anima \(Y\).

(2)

(Full faithfulness): For objects \(x,x' \in X_0\), the induced map \(f\colon \Hom _X(x,x') \to \Hom _Y(fx, fx')\) is an equivalence of animae.

Remark 25.1.6. Note that a functor \(F\colon C \to D\) between \(\infty \)-categories is essentially surjective (resp. fully faithful) precisely if the induced map on nerves \(N(F)\colon N(C) \to N(D)\) is essentially surjective (resp. fully faithful). In particular, \(F\) is an equivalence if and only if \(N(F)\) is a Dwyer-Kan equivalence.

Definition 25.1.7 (Complete Segal anima, Rezk (2001)). A Segal anima \(X\colon \simp \catop \to \An \) is called complete, or is said to satisfy the Rezk condition, if the map \[ X_0 \to X_1^{\sim }, \quad x \mapsto \id _x \] is an equivalence of animae. We denote the full subcategory of complete Segal animae by \(\CSeg (\An ) \subseteq \Seg (\An )\).

Taking fibers of the completeness equivalence over \((x,y)\in X_0\times X_0\) identifies the anima \((x=_{X_0}y)\) with the anima of isomorphisms from \(x\) to \(y\) in \(X\). Thus, for complete Segal animae, the two notions agree.

The nerve \(N(C)\) of an \(\infty \)-category is complete. In particular, the nerve functor \(N\colon \Cat _{\infty } \to \sAn \) factors through \(\CSeg (\An ) \subseteq \sAn \). This assertion is also included in Chapterexercise 25.2.

Lemma 25.1.8 ([Hebestreit and Steinebrunner (2025), Lemma 2.3]). Let \(\phi \colon X \to Y\) be a morphism of complete Segal animae. Then \(\phi \) is an equivalence if and only if it is a Dwyer-Kan equivalence.

Proof. See Corollary 15.2.13 of [Cisinski et al. (2026)]. □

Recall from Construction 1.8.12 the associated \(\infty \)-category functor \[ \ac \colon \sAn \to \Cat _{\infty }, \] which is the colimit-preserving extension of the inclusion \(\simp \hookrightarrow \Cat _{\infty }\). By Theorem 12.8.1 of [Cisinski et al. (2026)], this functor is left adjoint to the nerve functor: \[ \ac \colon \sAn \rightleftarrows \Cat _{\infty }\noloc N. \]

The pointwise formula for left Kan extensions (Theorem 22.4.3) expresses \(\ac (X)\) as a colimit in \(\Cat _{\infty }\). Using Proposition 24.5.2, this gives a description of \(\ac (X)\) as a localization; see [Hebestreit and Steinebrunner (2025), Corollary 3.8]. That description is used to prove the following result:

Proposition 25.1.9 (Hebestreit and Steinebrunner (2025), Corollary 3.15, Lemma 4.1). For every Segal anima \(X\), the unit map \(\eta \colon X \to N(\ac (X))\) is a Dwyer-Kan equivalence.

We refer to [Hebestreit and Steinebrunner (2025)] for the proof. For complete Segal animae, the unit is therefore an isomorphism, which will give the following theorem:

Theorem 25.1.10 ([Joyal and Tierney (2007), Section 4], [Lurie (2009a), Corollary 4.3.16], [Hebestreit and Steinebrunner (2025)]). The nerve functor \[ N\colon \Cat _{\infty } \hookrightarrow \sAn \] is fully faithful. Its image consists of the complete Segal animae. In particular, the nerve functor induces an equivalence \[ N\colon \Cat _{\infty } \iso \CSeg (\An ) \] between \(\Cat _{\infty }\) and the \(\infty \)-category of complete Segal animae.

Proof. Since the nerve functor lands in \(\CSeg (\An )\), the adjunction \(\ac \dashv N\) induces an adjunction

Commutative diagram involving accolon CSeg(An), \Cat infinity noloc N.

We have to show that both the unit and the counit are equivalences. For the unit, let \(X\) be a complete Segal anima. The unit map \(\eta \colon X \to N(\ac (X))\) is a Dwyer-Kan equivalence by Proposition 25.1.9 and thus an equivalence by Lemma 25.1.8, since both sides are complete Segal animae. For the counit, let \(C\) be an \(\infty \)-category. To show that \(\varepsilon \colon \ac (N(C)) \to C\) is an equivalence, it suffices by Remark 25.1.6 to show that \(N(\varepsilon ) \colon N(\ac (N(C))) \to N(C)\) is an equivalence. But this is a consequence of the triangle identity and the fact that \(\eta _{N(C)}\colon N(C) \to N(\ac (N(C)))\) is an equivalence. □

Corollary 25.1.11. The associated category functor \(\ac \colon \Seg (\An ) \to \Cat _{\infty }\) is a Bousfield localization. A morphism \(\phi \colon X \to Y\) of Segal animae is inverted by \(\ac (-)\) if and only if it is a Dwyer-Kan equivalence.

Proof. The first statement is a consequence of the full faithfulness of the nerve functor \(N\colon \Cat _{\infty } \hookrightarrow \Seg (\An )\) from Theorem 25.1.10. For the second statement, consider the following commutative square:

Commutative square with maps X to Y labeled phi, N(ac(X)) to N(ac(Y)) labeled N(ac(phi)), X to N(ac(X)) labeled \eta X, Y to N(ac(Y)) labeled \eta Y.

The vertical morphisms are Dwyer-Kan equivalences by Proposition 25.1.9. Since Dwyer-Kan equivalences satisfy the two-out-of-three property by Lemma 15.4.5, it follows that \(\phi \) is a Dwyer-Kan equivalence if and only if \(N(\ac (\phi ))\) is. By Remark 25.1.6, the latter holds if and only if \(\ac (\phi )\colon \ac (X) \to \ac (Y)\) is an equivalence. □

Corollary 25.1.12. The inclusions \(\CSeg (\An ) \hookrightarrow \sAn \) and \(\CSeg (\An ) \hookrightarrow \Seg (\An )\) admit left adjoints given by \(X \mapsto N(\ac (X))\).

Proof. Given a complete Segal anima \(Y\), we have \begin {align*} \Hom _{\CSeg (\An )}(N(\ac (X)),Y) &\simeq \Hom _{\Cat _{\infty }}(\ac (X),\ac (Y)) \\ &\simeq \Hom _{\sAn }(X,N(\ac (Y))) \\ &\simeq \Hom _{\sAn }(X,Y), \end {align*}

where the first isomorphism holds by the equivalence \(\Cat _{\infty } \simeq \CSeg (\An )\), the second comes from the adjunction \(\ac \dashv N\), and the third is induced by \(Y \iso N(\ac (Y))\). This proves the assertion for the inclusion into \(\sAn \); restricting to the full subcategory \(\Seg (\An ) \subseteq \sAn \) proves it for the second inclusion as well. □

Notation 25.1.13. The left adjoint to \(\CSeg (\An ) \hookrightarrow \Seg (\An )\) is often called the completion functor or Rezk completion and is denoted by \[ \widehat {(-)}\colon \Seg (\An ) \to \CSeg (\An ). \]

As a consequence of the theorem, every small \(\infty \)-category can be built from points and arrows by iterated colimit constructions:

Proposition 25.1.14. The \(\infty \)-category \(\Cat _{\infty }\) is generated under colimits by \([0]\) and \([1]\).

Proof. The presheaf category \(\sAn = \Fun (\simp \catop ,\An )\) is generated by the representable presheaves under colimits by Theorem 1.8.10. It follows that \(\Cat _{\infty }\) is generated under colimits by their images under the localization functor \(\ac \colon \sAn \to \Cat _{\infty }\), which are the posets \([n]\). The claim now follows, since the Segal condition shows that each \([n]\) is an iterated pushout of copies of \([1]\) along \([0]\). □